ECOLOGICAL RESEARCH IN NATIONAL
PARKS OF THE PACIFIC NORTHWEST
Compiled from Proceedings of the Second
Conference on Scientific Research in the National Parks
San Francisco, California
November 1979
Copies available from:
National Park Service Cooperative Park Studies Unit
Forestry Sciences Laboratory
3200 Jefferson Way
Corvallis, Oregon 97331
Ecological Research
in National Parks of the Pacific Northwest
An Oregon State University Forest Research Laboratory Publication
Corvallis, Oregon 97331
(Compiled from Proceedings of the Second Conference on Scientific Research
in the National Parks, San Francisco, California, November 1979)
TECHNICAL COORDINATORS
Edward E. Starkey, Project Leader
Cooperative Park Studies Unit
National Park Service
Oregon State University
Corvallis, Oregon 97331
Jerry F. Franklin, Project Leader
U.S. Forest Service
Pacific Northwest Forest and Range Experiment Station
Forestry Sciences Laboratory
3200 Jefferson Way
Corvallis, Oregon 97331
Jean W. Matthews, Science Editor
Cooperative Park Studies Unit
National Park Service
Forestry Sciences Laboratory
3200 Jefferson Way
Corvallis, Oregon 97331
Published September, 1982
Qhioh-s
. N k> C Co (d
Trade names and commercial products may be named
in the text only as a convenience to the reader;
but no endorsement by the U.S. Department of
Agriculture or the U.S. National Park Service is
implied.
FOREWORD
Three main points emerge from this collection of
papers... 1) the value of natural areas for research
into the workings of the natural systems, 2) the
fiscal and informational benefits of interagency
cooperation and cross disciplinary approaches to
research, and 3) the importance of such research
to managers of both manipulated and natural sites.
One of the most rewarding outcomes of this work has
been the "serendipitous" insights that specialists,
working together, have achieved. Specialization is
absolutely basic to thorough, in-depth science,
but the meshing of a number of disciplines in a
concerted effort to understand one particular site
or problem can result in unique flashes of insight.
Those insights can come through papers assembled
from various disciplines into a generally focused
volume, or they can arise from the meshing of minds
in an intensive interdisciplinary exercise like the
Hoh River Pulse Study (from which seven of these
papers were drawn).
Just as important as the excitement such sparks
generate among individual scientists is the useful
comparison the studies provide for managers of
natural areas and for managers of manipulated
areas. The store of information about how best to
perform each task is considerably enhanced by
findings of the kind presented here.
Natural areas, most notably National Parks,
Research Natural Areas, and Biosphere Reserves,
encompass a broad set of representative ecosystems
operating today as nearly as possible in the
pathways they would have taken if technology had
not become such a decisive factor in, and on, the
biosphere of Earth. To the extent that technology
is an outgrowth of natural systems, it must be
considered "natural." Volcanoes, however, are
natural also; and we recognize that they have had
devastating effects on the ecosystems in place
when they erupted. By bringing together in one
volume a cross section of scientific investigation
into the functioning of the underlying natural
systems that support our technology and our lives,
we are documenting the state of our scientific
investigation at this time. We are also demon-
strating the way scientists from two federal
agencies with different missions have worked
together to provide additional light on the
agencies' subject areas.
This volume then, represents a joint recognition
on the part of two caretaker agencies that natural
areas are the best remaining "libraries" for under-
standing the natural systems that brought us all
to our present position as the prime manipulators
in the history of Earth. It suggests that among
the most immediate future roles for these areas
may be as baseline "controls" against which to
measure similar environments being manipulated.
And finally, it provides a richer understanding of
the Northwest's ecosystems and of the future possi-
bilities for extending their uses, meanwhile pre-
serving their on-going health and productivity.
6404CC
CLEMSON UNIVERSITY LIBRARY
PREFACE
The National Park Service and the American Insti-
tute of Biological Sciences sponsored the Second
Conference on Scientific Research in National Parks,
November 26-30, 1979, in San Francisco, California.
More than 500 scientists and managers attended from
local, State, and Federal organizations.
Presented at the Conference were 20 papers dealing
with various aspects of the overall ecology of
National Parks and a National Monument in the
Pacific Northwest Region. These papers do not
cover the breadth of research in this area, but
are instead examples of current issues that
concern researchers and resource administrators.
The papers also reflect the growing base of
knowledge acquired as a result of shared effort
among a variety of research support agencies,
organizations, and institutions.
The USDA Forest Service, Environmental Protection
Agency, Army Corps of Engineers, and several
universities and private industries all provided
principal investigators and cooperation. Without
this support, most of the research reported herein
would not have been initiated nor could this
report have been published.
The papers are printed here essentially as
submitted by the authors except for some minor
copy editing to assure uniformity of style.
Jean Matthews
Science Editor
National Park Service
The diverse range of research topics covered could
only have been studied in relatively undisturbed
areas such as National Parks. Examples are the
papers dealing with climax forests, ungulate popu-
lations that are unhunted, and habitats ungrazed
by domestic livestock, where air and water are
clean enough to serve as a baseline for pollution
measurements .
CONTENTS
1 Ecosystem Studies in the Hoh River Drainage, Olympic National Park
Jerry F. Franklin
9 Habitat and Food Resources for Invertebrate Communities in South Fork Hoh River,
Olympic National Park
G. Milton Ward, Kenneth W. Cummins, Robert W. Speaker, Amelia K. Ward, Stanley V. Gregory,
and Thomas L. Dudley
15 The Biomass, Coverage, and Decay Rates of Dead Boles in Terrace Forests, South Fork Hoh River,
Olympic National Park
Robin Lee Lambert Graham
22 Structure, Composition, and Reproductive Behavior of Terrace Forests, South Fork Hoh River,
Olympic National Park
Arthur McKee, George LaRoi, and Jerry F. Franklin
30 Interactions Among Fluvial Processes, Forest Vegetation, and Aquatic Ecosystems, South Fork Hoh River,
Olympic National Park
Frederick J. Swanson and George W. Lienkaemper
35 Ecology and Habitat Requirements of Fish Populations in South Fork Hoh River, Olympic National Park
J. R. Sedell, P. A. Bisson, J. A. June, and R. W. Speaker
43 Relationships Within the Valley Floor Ecosystems in Western Olympic National Park: A Summary
Jerry F. Franklin, Frederick J. Swanson, and J. R. Sedell
46 Introduction and Dispersal of Mountain Goats in Olympic National Park
Bruce B. Moorhead and Victoria Stevens
51 Factors Reflecting Mountain Goat Condition and Habitat Quality: A Comparison of Sub-populations in
Olympic National Park
Victoria Stevens
58 Mother-Infant Interactions Among Free-Ranging, Non-Native Mountain Goats (Oreamnos americanus) in
Olympic National Park
Michael Hutchins and Craig Hansen
69 Home Range and Habitat Use by Non-Migratory Elk (Cervus elaphus roosevelti) in Olympic National Park
Kurt J. Jenkins and Edward E. Starkey
77 Pollutant Monitoring in the Olympic National Park Biosphere Reserve
Kenneth W. Brown and G. Bruce Wiersma
83 Research/Managment Prescribed Burning at Lava Beds National Monument
Robert E. Martin, Craig M. Olson, and James Sleznick, Jr.
92 Effects of Prescribed Fires on Vegetation in Lava Beds National Monument
Craig M. Olson, Arlen H. Johnson, and Robert E. Martin
101 Single-Year Response of Breeding Bird Populations to Fire in a
Curlleaf Mountainmahogany-Big Sagebrush Community
Tamara E. Tiagwad, Craig M. Olson, and Robert E. Martin
111 Effects of Prescribed Burning On Mule Deer in Lava Beds National Monument
Alice Purcell, Roger Schnoes, and Edward E. Starkey
121 Fire in the Forests of Mount Rainier National Park
Miles A. Hemstrom
127 Forest Dynamics and Fuelwood Supply of the Stehekin Valley, Washington
Bruce C. Larson and Chadwick Dearing Oliver
135 Optical Properties of Crater Lake, Oregon: Variation in Secchi Disk Transparency, 1937-79
Douglas W. Larson and Mark E. Forbes
138 Species Composition and Vertical Distribution of Pelagic Zone Phytoplankton in
Crater Lake, Oregon: 1940-79
Douglas W. Larson and N. Stan Geiger
Digitized by the Internet Archive
in 2013
http://archive.org/details/ecologicalresearOOunse
Ecosystem Studies in the Hon River Drainage, Olympic National Park
Jerry F. Franklin
Jerry F. Franklin, U.S. Department of Agriculture,
Forest Service, Pacific Northwest Forest and Range
Experiment Station, Forestry Sciences Laboratory,
3200 Jefferson Way, Corvallis, Oregon 97331.
Interdisciplinary, ecosystem-oriented research is
essential to understanding complex, interlinked
resource values. A project of this type was
carried out in the South Fork of the Hoh River in
the western Olympic Mountains. This wilderness
drainage has outstanding examples of broad
terraces with Picea sitchensis-Tsuga heterophylla
rainforests, a major glacial-fed river, and
fluvial processes. During an intense 11-day
"pulse," relationships were examined between
geomorphic processes, terrestrial communities, and
aquatic systems. This paper discusses the
rationale of the study, describes the objectives
and approaches, and serves as introduction to
seven papers that follow.
On September 5, 1978, a large group of scientists,
technicians, and graduate students entered the
South Fork of the Hoh River drainage. The
research team included personnel from 11 different
organizations and a broad range of disciplines.
During the next 11 days, 46 technical personnel
devoted 266 field days to an examination of
landforms, geomorphic processes, and aquatic and
terrestrial communities and their interrela-
tionships.
My objective in this paper is to outline the
rationale and objectives of this short-term
intensive research effort which we refer to as a
"pulse." I will also describe the essential
features of the site. All of this serves as
introduction to the seven papers that follow. Six
of these papers are based on research conducted in
the South Fork during the pulse. One paper (see
Jenkins and Starkey in this report), based on work
in the main Hoh River drainage, is included in
this session because of the importance of the
Roosevelt elk herds in these valley ecosystems.
RATIONALE
It is increasingly apparent that integrated
studies of natural ecosystems are critical to
solution of management problems as well as to the
general progress of natural history research.
Projects like the Coniferous Forest Biome have
demonstrated both the practicality and value of
integrated examinations of natural ecosystems
(Edmonds 1980). In such studies emphasis is
placed on linkages between ecosystem compartments,
such as the interface between a forest and stream;
linkages which are often avoided or not considered
in traditional disciplinary research. There is
increasing evidence of the necessity for focusing
on linkages with many examples of difficulties
encountered in considering problems in isolation,
whether it be herbivores viewed outside of a
habitat context or a stream analysis that fails to
consider terrestrial inputs. Land managers are
increasingly faced with problems involving
multiple linkages and resource tradeoffs; and
their needs often can only be met with integrated,
ecosystem-level research.
A corps of interested scientists and associates
with a tradition of integrated, ecosystem-oriented
research has developed around programs centered at
Corvallis, Oregon. Personnel include staff of
Oregon State University and the USDA Forest
Service Forestry Sciences Laboratory as well as
"graduates" who have moved out into industry and
other agencies. Joint research efforts are an
essential factor in maintaining the interdisci-
plinary linkages and system-level perspective
among the corps. Hence, the desire and need for
periodic field exercises.
The south fork of the Hoh River represents a
complete river drainage from 20 km from its
headwaters to the Olympic National Park boundary.
The drainage is essentially pristine with very
light recreational use, one short trail, and no
roads. Although the river is relatively small,
fluvial processes are evident; and the valley
bottom is broad with extensive alluvial land-
forms. Geomorphic-biologlc interactions can,
therefore, be viewed within a relatively compact
but ecologically complete river drainage.
Baseline data are needed to serve managerial and
scientific purposes on ecosystems within both
Olympic National Park and the South Fork drainage.
Basic resource information is, of course, welcomed
by Park managers and interpreters. Management-
related data on the fisheries resource are
important in Olympic National Park. Recent
designation of Olympic National Park as a
Biosphere Reserve increases the need for expanded
data bases and monitoring programs. Finally, the
South Fork of the Hoh River is identified as a
potential Research Natural Area because of the
outstanding valley-bottom Picea sitchensls forests
and need for further evaluation.
There also is a basic need for scientific data on
ecosystems of the type found in the South Fork
drainage. Substantial research exists on forest
and stream interactions, but almost none has been
done on northwestern rivers. Information is
needed on natural fisheries and sediment levels in
a pristine but natural sediment-rich river system;
these data provide a baseline for comparison with
adjacent Olympic Peninsula river systems that are
being logged. Further, basic knowledge of western
Olympic Mountain valley bottom forests (Fonda
1974) is still sparse and includes essentially
little on population structure of the trees and on
coarse woody debris.
Thus, a variety of factors combined to make
desirable an integrated research project on the
South Fork of the Hoh River. An interdisciplinary
examination of Olympic rainforests and associated
streams and rivers was needed. Appropriate
methodology and perspectives were available along
with a corps of personnel with a suitable mix of
disciplines. An outstanding site existed in the
South Fork drainage. Baseline data and permanent
sample plots were generally needed for the National
Park/Biosphere Reserve, and managers had current
needs for specific types of data. Research in
these coastal forests was also needed to advance
basic ecosystem science in the Pacific Northwest.
STUDY AREA
The drainage of the South Fork, of the Hoh River is
located on the western slopes of the Olympic Moun-
tains at about 47°47' N. latitude and 123°56' W.
longitude (fig. 1). The South Fork is a glacial
river arising from Hubert Glacier on the slopes of
Mount Olympus and running for about 25 km to its
confluence with the main fork of the Hoh River.
Most of the drainage is located within Olympic
National Park (fig. 2).
The South Fork drainage covers about 11,400 ha
within the Park and is generally a broad,
glacially carved, u-shaped valley. Adjacent
mountain slopes are precipitous and composed
largely of sandstones and shales. The floodplain
in the lower valley is exceptionally wide,
occupying nearly 25 percent of the total width of
the valley at the primary study sites; the
floodplain in the main fork of the Hoh River is,
by contrast, only 10 percent of the width of the
valley. The landforms in the lower valley are
discussed by Swanson and Lienkaemper in this
report .
Climatically, the study area is extremely wet and
mild. The Spruce Weather Station is located along
the Hoh River below the study area (U.S. Department
of Commerce 1965). Precipitation there averages
over 3 200 mm annually with 55 mm in the driest
month (July). Snow is uncommon. Mean temperatures
are probably around 10°C with January minima of
around 1°C and July maxima of 21 °C. Fog and low
clouds often occupy the valley even when higher
mountain areas are experiencing clear weather.
Precipitation occurred every day between September
5 and 15, 1979, and totalled at least 200 mm.
Rains in excess of 100 mm at the camp during a
2-day period resulted in water ponding on the
lower terrace and a substantial increase in the
height of the river.
Environmental stresses are obviously uncommon in
these valley bottom forests. Snow, ice, and
drought play little or no role. Windstorms do
cause significant tree mortality, particularly
from strong southwesterly winds associated with
major winter storms. Catastrophic blowdowns do
occur every few decades in localized areas. A
major blowdown actually occurred in January 1979
and caused significant mortality of mature trees
within the permanent sample plots. Fire appears
to be an insignificant factor on the terraces
themselves; the only charcoal found in soil pits
appeared to have been transported to the site.
Fire has been an important factor on the mountain
slopes.,
The research was confined primarily to the valley
bottom environment with very little sampling of
the mountain slope or river headwater environments.
The major study sites are located 3 to 5 km
upstream from the National Park boundary at about
215-m elevation. Forest sampling was confined to
terraces and river bars except for anchor points
of the two longest transects which were located on
mountain slopes. Terrace habitats were generally
not confounded with colluvial and alluvial
• OLYMPIA
Figure 1. — Olympic Peninsula in Washington State
showing general location of Olympic National Park
and the Hoh River drainage; coastal strip of the
National Park not shown.
deposits from sideslopes and alluvial fans from
tributary drainages as is the case along much of
the main Hoh River.
LOGISTICS
The logistical arrangements were developed jointly
by Olympic National Park and the research team
leaders. Base camp was located 3.5 km upstream
from the Park boundary on a river bar in order to
minimize long-term impacts of a large group on the
valley. Equipment and supplies necessary for base
camp and the research were brought in by heli-
copter. Research personnel brought in their own
gear over 5 km of trail from the road head.
A total of 46 persons contributed at least 1 day
of field work. Organizations represented in the
group included the National Park Service, USDA
Forest Service, Oregon State University,
Weyerhaeuser Company, University of Washington,
Washington Department of Natural Resources,
University of Alberta, University of Edinburgh,
U.S. Geological Survey, and U.S. Fish and Wildlife
Service. Not all personnel were present on any
single day; average daily participation was 26,
not including visitors. Teams were formed to do
individual tasks with personnel leaving or being
reassigned to new tasks upon completion of an
activity.
Boundary,
Olympic Natl. Park
Mount Olympus
A.
Boundary,
South Fork Hon River drainage
10
Kilometers
Figure 2. — The South Fork Hoh River drainage showing location of major study site.
A key feature of the
review session during
and discussed that da
were critical in modi
research effort as we
ciplinary exchange of
phenomena of common i
sometimes resulted in
A large team project
semi-structured excha
desired communication
pulse was the regular evening
which personnel presented
y's findings. These sessions
fying and sharpening the
11 as insuring an interdis-
viewpoints on features and
nterest. Such exchanges
serendipitous discoveries,
needs to provide for
nges of this type to get the
and collaboration.
OBJECTIVES AND APPROACHES
The major objectives of the studies in the South
Fork of the Hoh River are outlined in table 1.
Many of the tasks appear relatively independent
although all relate to the overall objective of
describing and better understanding the valley
bottom ecosystems found on the western slopes of
the Olympic Mountains. Interdisciplinary efforts
often break down into component tasks, some of
which are strongly discipline oriented. The
overall design provides the context which makes
the parts fit into a whole. In fact, in the Hoh
studies each objective is linked to at least one
other objective, generally by a requirement for
information. Several objectives, such as the
definition of aquatic habitats and their relation
to geomorphic processes and terrestrial vegeta-
tion, obviously require interdisciplinary
collection and synthesis of data.
The basic geomorphic analysis of the valley bottom
landforms and processes is covered by Objective I
(table 1). Fonda (1974) had developed a model of
vegetation-landf orm relationships for the main
fork of the Hoh River, and the research team
wanted to test its application in the South Fork.
Special interest centered on interactions of
landforms, and geomorphic processes with terres-
trial vegetation and major reciprocal effects do
exist (see Swanson and Lienkaemper in this report).
Geomorphic processes and vegetation also link to
produce certain aquatic habitats. The major
approaches to Objective I were mapping of
elevation-vegetation profiles along permanent
transects laid out across the valley floor (see
Swanson and Lienkaemper in this report) .
Objectives II, III, and IV focus on descriptions
of the valley bottom forests with special
attention to a key structural component (dead
wood) and the reproductive population dynamics of
the two major tree species — Picea sitchensis and
Tsuga heterophylla. The research team was
interested in obtaining data on amounts and
decomposition rates of coarse woody debris in
coastal environments to compare with a large data
base collected from Cascadian conifer forests.
The team also hypothesized that reproductive
behavior would be influenced by down logs although
the overwhelming importance of logs (see McKee et
al. in this report) was not appreciated at the
outset of the study. The relative reproductive
success and presumed ecological role of Picea and
Tsuga in the Olympic rainforests has been an
unresolved topic of discussion (Franklin and
Dyrness 1974, Fonda 1974).
Since Objectives II, III, and IV (table 1) gen-
erally utilized the same data base, the sampling
techniques are detailed here to avoid repetition.
Sampling of the forest communities began with
reconnaissance and systematic sampling at inter-
vals along four line transects run perpendicular
to the river and across several landforms. The
line transects were also used by the geomorphic
research team. The transect data are incorporated
into the general compositional and structural
descriptions of the valley forests in the study
area (see McKee et al. in this report).
Reconnaissance and the transect sampling revealed
two distinct, mature forest communities of Picea
sitchensis and Tsuga heterophylla. These are
described in detail by McKee et al. in this report
but generally consist of an open Picea sitchensis/
Acer c ire ina turn/herb forest on lower terrace
surfaces (fig. 3) and Picea sitchensis-Tsuga
heterophylla/Vaccinium sp./moss forest on high or
upper terraces (fig. 4).
Permanent hectare (100- x 100-m) sample plots were
established in typical segments of these two for-
est types to facilitate sampling of coarse woody
debris and tree regeneration and collection of
additional compositional and structural data on
mature forests. Two continuous plots were estab-
lished in each forest for a total of 4 hectares.
The basic layout of the plots followed the proce-
dures developed for reference stands at the H. J.
Andrews Experimental Forest (Hawk et al. 1979).
Plot boundaries were surveyed with staff compass
and tape. Plots were then temporarily gridded
with string into segments as small as 5 m to ea=^
ocular mapping of all live trees 2 5 cm, snags s.
15 cm, and logs > 10 cm in diameter. Since the.^e
were permanent plots, all live trees over 15 cm
diameter were tagged with numbered metal tags at
breast height. Galvanized steel pipe approximately
1.5 m in height and 4 cm in diameter were installed
at the corners, center, and quarter corners of
each plot. Live trees and snags were subsequently
Table 1 — Major objectives of South Fork Hoh River
ecosystem studies
I. Describe the valley landforms with some
specific interests
A. Role of vegetation in landform
development
B. Formation of different aquatic habitats
II. Develop baseline descriptions of the
valley-bottom forest including
A. Live, standing dead, and down trees
B. Relationships of forest types to landform
C. Descriptions developed as a permanent
sampling system
III. Analyze the amounts and role of dead and
down wood in the valley-bottom forests
IV. Analyze the regeneration dynamics of forest
trees in the valley-bottom forests
V. Describe and analyze aquatic habitats in the
lower valley
A. Define aquatic habitats and determine
their relation to geomorphic and
terrestrial processes
B. Determine biology of habitats, energy
base, and invertebrate communities
C. Baseline data on sediments
VI. Analyze use of aquatic habitats by fish
A. Species distributions
B. Habitat use by anadromous fish
C. Overall importance to total fishery
VII. Examine interactions between Roosevelt elk
and vegetation
Figure 3. — Mature forests on the lower terrace are
typified by large, widely spaced Picea sitchen-
sis, a tall shrub layer of Acer circinatum, and a
dense herbaceous ground layer that is rich in
grasses.
Figure 4. — Well-stocked stands of Picea sitchensis
and Tsuga heterophylla typify mature forests on
upper terraces along with understories dominated
by Vaccinium sp., ferns, and mosses.
measured for height. Areas of standing water and
elk trails also were mapped. The resulting stand
maps for two of the hectare plots are shown in
figures 5 and 6. Additional sampling of down logs
(see Graham in this report) and regeneration (see
McKee et al. in this report) was done on these
permanent plots.
Objectives of the aquatic research involved
definition of distinct aquatic habitats with the
assistance of the geomorphologists followed by a
thorough characterization of their biology — energy
sources, invertebrate communities, and type and
level of usage by various fish (table 1). The
habitat classification scheme is discussed by
Swanson and Lienkaemper and Ward and Cummins both
in this report. The aquatic biologists and
fisheries researchers selected representative
areas of each habitat for their sampling (see Ward
and Cummins and Sedell et al. in this report).
The research on sediments and anadromous fish also
involved extensive sampling along nearly 10 km of
the South Fork.
The objective (VII) on Roosevelt elk-vegetation
interactions is listed last because little was
accomplished as part of the South Fork pulse; not
because it is unimportant. Roosevelt elk are a
significant component of these ecosystems and
several hypotheses have been proposed about their
effects on plant composition and tree reproduction
(see McKee et al. in this report). We are
grateful that Jenkins and Starkey agreed to
include their paper in this report, which is based
on research In the main Hoh River valley, with
this series from the South Fork. Dr. D. Boersma,
of the Environmental Research Institute of the
University of Washington, has initiated a study on
effects of elk grazing on tree reproduction. This
research, along with the planned establishment of
0. 5-ha exclosures around portions of the permanent
sample plots on both the upper and lower terraces,
should begin providing some quantitative
information on elk-vegetation interactions.
CONCLUSION
This paper introduces and places in perspective a
series of seven papers on valley-bottom ecosystems
in the Hoh River drainage. The concluding paper
(see Franklin et al. in this report) is a brief
summary emphasizing the major conclusions and
reiterating the interrelationships between
ecosystem components.
LITERATURE CITATIONS
Edmonds, Robert L. , editor.
1980. The natural behavior and response to
stress of western coniferous forests. US/IBP
Synthesis Series. Dowden, Hutchinson & Ross,
Inc., Stroudsburg, PA. In Press.
Fonda, R. W.
1974. Forest succession in relation to river
terrace development in Olympic National Park,
Washington. Ecology 55(5) :927-942, illus.
Franklin, Jerry F. , and C. T. Dyrness.
1973. Natural vegetation of Oregon and
Washington. USDA Forest Service Gen. Tech. Rep.
PNW-8, 417 p., illus. Pac. Northwest For. and
Range Exp. Stn. , Portland, OR.
Hawk, G. M., J. F. Franklin, W. A. McKee, and
W. B. Brown.
1979. H. J. Andrews Experimental Forest
reference stand system: establishment and use.
Coniferous Forest Biome Bull. 12, 79 p., illus.
Univ. of Wash. Seattle, WA.
U.S. Department of Commerce.
1965. Climatic summary of the United States-
supplement for 1951 through 1960. Washington,
D.C. Climatography of the United States No.
86-39. 92 p.
SOUTH FORK HOH UPPER TERRACE
PLOT 1
9- PSEUDOTSUGA MENZIESII
O TSUGA HETEROPHYLLA
O PICEA SITCHENSIS
€ THUJA PLICATA
• ACER CIRCINATUM
V^LOG
o I5-25CM
o 25-50 CM
O 50-IOOCM
O I00-200 CM
O > 200 CM
O STUMP
* POST
N75°E
Figure 5. — Stem map of permanent sample plot 1
located on the upper terrace.
SOUTH FORK HOH LOWER TERRACE
PLOT 4
TREE SIZE CLASS
o < 25 CM
O 26-50 CM
O 51 -100 CM
O 100-150 CM
O > 150 CM
TREE SPECIES
O PICEA SITCHENSIS
O TSUGA HETEROPHYLLA
• ACER CIRCINATUM
© ALNUS RUBRA
€) THUJA PLICATA
ELK TRAILS
N22°E
Figure 6. — Stem map of permanent sample plot 4
located on the lower terrace; note the reduced
density of logs and stems and the extent of
standing water.
LOGS
PROMINENT ELK TRAIL -
DECAY CLASS I- DC EXPOSED MINERAL SOIL
^?H STANDING DEAD INTERMEDIATE -ELK TRAIL- COMB
OF EXPOSED S0IL + INTACT VEG
INCIPIENT TRAIL, MINERAL SOIL
NOT EXPOSED, INTACT VEG.
X LOG DEEPLY CUT BY ELK CROSSING
A LOG MARKED BY ELK CR0S9NG-N0T DEEPLY CUT
STANDING
WATER
Habitat and Food Resources for Invertebrate Communities in South Fork
Hoh River, Olympic National Park, Washington
G. Milton Ward, Kenneth W. Cummins, Robert W. Speaker,
Amelia K. Ward, Stanley V. Gregory, and Thomas L. Dudley
ABSTRACT
The morphological and biological structure of four aquatic habitats in the valley of the South Fork Hoh
River are described, and the potential effect of naturally occurring inorganic sediments (glacial flour)
is discussed. The main river habitat was heavily impacted by the hydrologic regime of the river, com-
pounded by the scouring action of a large suspended inorganic sediment load. River off-channel and terrace
tributary habitats were less affected by these events, and they developed relatively larger invertebrate
communities. Small, high gradient streams draining valley side slopes were not affected by main river
processes but possessed a potentially flashy hydrologic regime. River off-channel and terrace tributary
habitats seemed to provide the optimal aquatic habitat in this river system.
G. Milton Ward, Kenneth W. Cummins, Robert W.
Speaker, Amelia K. Ward, Department of Fisheries
and Wildlife, Oregon State University, Corvallis,
Oregon 97331; Stanley V. Gregory, U.S. Fish and
Wildlife Service, Columbia National Fishery Re-
search Laboratory, Field Research Station, Oregon
State University, Corvallis, Oregon 97331; and
Thomas L. Dudley, Department of Entomology, Oregon
State University, Corvallis, Oregon 97331.
INTRODUCTION
For the few remaining pristine watersheds in the
Pacific Northwest, there exists a need for documen-
tation of the types of aquatic habitats present
and a description of their physical and biological
structure. An example of one such watershed is
the South Fork Hoh River, Olympic National Park,
Washington. This system was of interest because
of its size, relatively undisturbed state, and the
presence of large quantities of naturally
occurring inorganic sediment. Finely ground rock,
generated by glacial activity in the headwaters,
was found throughout the main river channel, back-
water areas, and in the downstream reaches of many
tributaries.
As a result of current grazing and forestry
practices in the Western United States, erosion of
valley slopes and stream banks has been greatly
accelerated in recent years.
Descriptive data on the South Fork Hoh River
system, with its natural, long-term sedimentation
patterns may be useful in assessing the impacts of
recent sedimentation in other streams.
The interplay of geomorphic processes and terres-
trial vegetation in the South Fork basin has
created four major aquatic habitats. Two of
these, the main river and river off-channel areas,
lie in the lowest section of the river valley at
or below the level of winter base flows. The two
other habitats are tributary systems, one that
drains steep valley walls and a second that flows
across the elevated terraces occurring along the
north side of the river.
The objective of this investigation was to
establish physical and biological descriptions of
the aquatic habitats in the South Fork Hoh River
Valley and to examine the influence of naturally
occurring inorganic sediment on these habitats.
Two specific objectives were communities and the
food resources available to them.
Invertebrate communities were examined from both a
taxonomic and ecological standpoint. Based on
insect feeding habits, Cummins (1974) classified
aquatic insect groups according to their
ecological role in stream ecosystems. Within a
system, four basic groups of organisms could be
recognized: species that feed on fine particle
detritus; scrapers, whose main diet is periphyton;
shredders, who feed primarily on coarse
particulate organic matter; and predators, who
feed on organisms in other functional groups.
Insect species composition in stream habitats vary
geographically; but in all areas, representa-
tives of these ecological groups can be found.
Figure 1. — River channel and off-channel habitats
in South Fork Hoh River, Olympic National Park,
Washington.
DESCRIPTION OF AQUATIC HABITATS AND
INVERTEBRATE COMMUNITIES
River Channel
Fed by cold glacial headwater streams, the main
stem of the South Fork Hoh River flows along a
wide, shallow, and unstable channel containing
rocks, small boulders, and considerable easily
transported inorganic matter (fig. 1). Much of
this fine inorganic sediment is generated by
glacial activity in the headwaters and then
transported in high concentrations throughout the
downstream reaches, even into terrace tributaries
during major flood events. This sediment (glacial
flour), which imparts a milky color to the water
when suspended and a greyish-white color to the
benthos, was very prevalent in river channel,
off-channel, and terrace tributary habitats, where
it often completely fills intergravel spaces along
stream margins.
For many other river channels equal in size to
that of the South Fork Hoh River, primary
production by attached algae is a major source of
energy and carbon for the system; however, here
that energy source was greatly reduced, being only
10 percent as great as has been measured in a
comparable system, the McKenzie River, Oregon
(Naiman and Sedell 1980) (table 1). The large
amounts of suspended inorganic sediment reduce
productivity by scouring algal cells from the rock
surfaces, and deposited sediments may bury many
potential algal sites. The large size of the main
channel did not allow terrestrially produced
energy sources to play a prominent role, and the
movement and deposition of the glacial flour
prevented much autochthonous production. As a
result, the system possessed a meager food base,
with little organic detritus and low algal
production.
The resulting invertebrate community in the main
channel was of generally low density and biomass
(tables 2 and 3). Fine particle feeding detriti-
vores, composed mostly of two mayflies, Baetis and
Rithrogena, were the dominant organisms (table
A). Algavores, which might typically inhabit the
upper surfaces of rocks in streams of this size,
were noticably missing, perhaps due to constant
scouring action of suspended glacial flour.
10
+
6
++
34
+++
12
1 1 1 1
11
Table 1 — Organic matter content in aquatic habitats, South Fork Hoh River
Allochthonous inputs:
Habitat ' TOM BOM Algal
mg/m-' % org g/m2 % org leaf litter wood chla
River channel 116 8 13 2
River off-channel — — 15 2
Terrace
tributaries 106 14 37 5
Valley wall
tributaries 209 33 21 12
TOM = Transported Organic Matter, BOM = Benthic Organic Matter.
Table 2 — Density of insect functional groups on aquatic habitats, South Fork Hoh River
Sediments (///m2 of rock) Leaf litter (///g)
Number ^
Habitat of taxa
Sh Col Scr Pred Total Sh Col Scr Pred Total
River channel 22 22 390 0 75 487 0.65 3.79 0 0.06 4.50
River off-channel 15 8 202 0 16 228 5.81 8.51 0 .20 14.52
Terrace tributaries 26 38 395 67 59 559 4.83 2.56 0 .17 7.56
Valley wall tributaries 23 13 411 182 20 626 8.87 6.21 .30 .30 15.68
Sh = Shredders, Col = Collectors/gatherers, Scr = Scrapers/grazers, Pred = Predators.
Table 3 — Biomass of insect functional groups in aquatic habitats, South Fork Hoh River
Habitat
Sediments (mg/m2 of rock) Leaf litter (mg/g)
Sh Col Scr Pred Total Sh Col Scr Pred Total
River channel 5 114 0 100 219 .53 2.18 0 .05 2.76
River off-channel 3 365 0 653 1022 2.23 3.63 0 .10 5.96
Terrace tributaries 9 301 84 35 429 4.45 2.05 0 .29 6.79
Valley wall tributaries <1 70 48 11 130 1.94 1.93 0 .20 4.07
Sh = Shredders, Col = Collectors/gatherers, Scr = Scrapers/grazers, Pred = Predators.
11
Table 4 — Taxonomic distribution of the most abundant invertebrates in
aquatic habitats, South Fork Hoh River
Habitat
Sediments (lt/m2)
Leaf litter (#/g)
River channel
River off-channel
Terrace tributaries
Valley wall tributaries
Baetis
164
Alloperla
.89
Ri throgena
71
Baetis
.71
Ecclisomyia
.53
Baetis
54
Capnia
3.11
Alloperla
54
Nemoura
2.49
Ecclisomyia
2.28
Baetis
59
Ecclisocosmoecus
2.31
Alloperla
46
Nemoura
1.94
Chronomidae
193
Alloperla
.87
Glossosoma
162
Nemoura
3.87
Limnephilidae
88
Epeorus
2.38
Chironomidae
175
Peltoperla
1.49
Baetis
81
Lepidostoma
1.49
River Off -Channel
A second habitat, lying within the geomorphically
active portion of the river channel (Surface 2,
Swanson and Lienkaemper in this report) was
comprised of areas previously part of the main
river channel. Because of deposition behind large
debris accumulations or because of the shifting
river bed, parts of the river were isolated
(fig. 1). Although they may carry water during
summer storm flows and during winter base flows,
these secondary and tertiary channels, comprising
the major portion of off-channel habitats, are
isolated from direct contact with the main river
channel during low flow periods. A connection
between the main river and these quiet off-channel
pools can be maintained, however, through inter-
gravel flow. This subsurface flow carries little
suspended glacial silt, since the gravel acted as
an efficient filter. Large amounts of glacial
flour were present in the sediments of these areas
(table 1), however, supporting the idea that river
and off-channel habitats often are connected.
During periods of moderately high flows when main
river and off-channel habitats are connected,
large debris accumulations are important in
protecting back-water habitats from severe
scouring (fig.l). The reduced flow rates in
protected areas create drop zones, however, where
suspended particles collect and settle out.
The organic energy inputs
to be greater than those
channel. There is greate
litter and wood input due
between the habitat and t
of our observations, thes
relatively large standing
chlorophyll a (table 1).
channel, standing crops o
very low; and the percent
sediments of these habita
to this habitat appeared
to the main river
r potential for leaf
to increased contact
he land; and at the time
e habitats exhibited
crops of algal
As with the main river
f organic detritus were
of glacial silt in the
ts was very high.
In spite of the potential impacts of the main
river during high discharges, these habitats
supported sizable invertebrate communities. The
predominantly inorganic sediments demonstrated low
insect densities but a high total bioraass
comprised mostly of predacious species (tables 2
and 3). Insects inhabiting leaf litter were
relatively abundant, the shredders Capnia, Nemoura
and Ecclisomyia among the more numerous (table 4).
Terrace Tributaries
Upon reaching the flat river terraces, valley wall
tributaries encounter a significant change in
slope. Water from these streams spreads out over
the terrace surface, often standing several inches
deep over a wide area. It seems likely that much
of this water percolates into the porous terrace
soil and reappears later as spring and seep
water. The flow in these terrace streams is
derived from valley wall tributaries and terrace
springs, as well as from the main river during
times of extremely high discharge (fig. 2).
These clear-water, sand- and gravel-bottomed
streams typically had a low suspended inorganic
load; benthic areas contained a relatively large
amount of organic detritus. Sediments still
contain substantial amounts of glacial flour,
reflecting occasional contact with the main river.
The nearly closed canopy and the presence of
shrubs along the stream margins made the potential
for allochthonous inputs quite high. Large woody
debris also was a common feature of this habitat.
The potential for algal production, while not as
high as in the off-channel habitats, was inter-
mediate between the main river and the off-channel
habitats. A number of stream reaches had very
high overstory canopies, and were well lighted.
The potential for algal activity in these areas
was quite good.
Invertebrate communities in the habitats were well
developed, especially so in the gravel bottom
areas, and exhibited wide taxonomic and functional
12
diversity (tables 2, 3, and 4). Terrace tribu-
taries appeared to offer a stable and relatively
productive area for aquatic invertebrate activity.
Due to the considerable spring activity, flow
rates were relatively stable and the trauma for
insect communities that might be associated with a
large suspended inorganic load was relatively
infrequent. It would seem that this was a very
important aquatic habitat in the Hoh River valley.
Valley Wall Tributaries
Streams draining valley side slopes represent the
fourth aquatic habitat examined in the South Fork
Hoh River basin (fig. 3). These habitats, like
many similar first- and second-order streams
throughout the western Cascades, were character-
ized by steep gradients, large accumulations of
woody debris, and heavy canopy cover. Average
channel slopes were large, but much of the drop
occurs in short steep falls over wood debris and
rocks. Downstream reaches of these steams flow
out into the river terraces, where water may stand
for a time before percolating into the soil.
These habitats appear to be heterotrophic, the
vast majority of the energy inputs comprised of
wood debris and coniferous needle litter (Sedell
and Triska 1975). Suspended loads and sediments
in the habitats contain relatively little glacial
flour, and total organic loading is the highest of
any of the four habitats examined (table 1).
Algal densities, as measured by chlorophyll a
concentrations (table 1), were typical of similar
habitats in the Cascades.
Densities of invertebrates in both benthos and
leaf litter were high, although biomass was not
(tables 2 and 3). Despite the heterotrophic
nature of this habitat, a relatively large
population of Glossosoma , a scraper, was present.
All functional groups of aquatic insects were
represented, including those species associated
with woody debris. Particularly abundant species
in the sediments and leaf litter were lepidosto-
matid and limnephilid caddisflies, as well as
neumourid and peltoperlid stoneflies and the
mayfly Epeorus (table 4).
CONCLUSIONS
Of the four aquatic habitats examined in the South
Fork Hoh River, the off-channel and terrace
tributary habitats were the most favorable for
aquatic invertebrates. Algal and invertebrate
communities were not as affected by suspended
glacial flour as were the river channel and valley
wall tributary communities, nor were they as often
subjected to hydrologic events capable of rearrang-
ing and destroying habitat. Food resources, too,
were more available. The major organic inputs to
river off-channel and terrace tributary habitats
were leaf litter and algae; inputs to the river
channel were extremely low and those to the valley
wall tributaries were comprised mainly of
refractory woody debris. The biomass of inverte-
brates in the terrace tributary and river
off-channel areas also was higher than in the
other two habitats. Total biomass and the groups
of organisms present reflected both the increased
physical stability of the habitats and the
continual inputs of high quality food resources.
Presence of glacial flour was noted in terrrace
tributary habitats as well as in river channel and
Figure 2. — Schematic drawing of terrace tributary
systems in valley of South Fork Hoh River.
Figure 3. — Debris map of a small Cascade mountain
stream very similar in physical and biological
structure to valley wall tributaries in the South
Fork Hoh River.
WATER FLOW
C-^T LOGl HT ABOVE LOW WATER, (M)
£= ROOT WAD
FLOATED ORGANIC DEBRIS
TRAPPED SEDIMENT
V-@ MINIMUM TIME AT SITE, YR
Q LIVING TREE
CC^J BOULDER ISLAND
CHANNEL BOUNDARY
13
off-channel areas. At times of extremely high
water, even the elevated terraces were inundated
by flood waters. Glacial flour was detected both
in the sediment and suspended in the water column;
but effects of suspended sediment seemed to be
greatest, particularly, in the main river
channel. In the absence of suspended sediment, a
river the size of the South Fork Hoh should be
very productive, yet the main river channel
habitat appeared to be quite the opposite. It is
likely that wherever suspended and deposited
inorganic sediments are comparatively high, stream
communities change in response to their presence;
however, those habitats which maintain a natural
food base and a stable relationship with the
terrestrial environment will maintain a diverse
and stable invertebrate community.
LITERATURE CITATIONS
Cummins, K. W.
1974. Structure and function of stream
ecosystems. Bioscience 24:631-641.
Naiman, R. J., and J. R. Sedell.
1980. Relationships between metabolic
parameters and stream order in Oregon. Can. J.
Fish. Aquat. Sci. 37:834-847.
Sedell, J. R. , and F. J. Triska.
1975. Biological consequences of large organic
debris in northwest streams. 1st Debris in
Streams Workshop, September 1975, Oregon State
University, 10 p.
H
The Biomass, Coverage, and Decay Rates of Dead Boles in Terrace Forests,
South Fork Hoh River, Olympic National Park
Robin Lee Lambert Graham
ABSTRACT
A two part study examining: (1) the quantity and spatial distribution of dead boles, and (2) the decay
rate of fallen boles by species and diameter was done in the terrace forests along the South Fork Hoh
River in the Olympic National Park. On the upper terrace, dead boles occupied 11 percent of the forest
floor and accounted for 165 tonnes/ha. The exponential decay rate for Sitka spruce was 0.0107 yr~ , for
western hemlock it was 0.0124 yr--*-. In both species, smaller diameter boles decayed more rapidly.
Robin Lee Lambert Graham, Department of Botany and
Plant Pathology, Oregon State University,
Corvallis, Oregon 97331.
15
The importance and ecosystem function of large
woody debris only recently has been explored
(Grier 1978, Franklin and Waring 1980, Lambert et
al. 1981). In forest ecosystems in the Pacific
Northwest, where dead boles typically account for
200 tonnes/ha of biomass and occasionally as much
as 500 tonnes/ha, they represent an enormous pool
of carbon, nitrogen, and mineral elements
(Franklin and Waring 1980). In addition, these
dead boles may persist for centuries; Douglas-fir
typically lasts over 300 years. Because of their
longevity and massive quantity, dead boles have
been hypothesized to function as ecosystem
stabilizers carrying the nutrients and stored
energy of an ecosystem through disturbances
(O'Neill et al. 1975).
Not only does abundant dead wood persist for long
periods of time, it also accumulates nitrogen,
phosphorous, and sometimes calcium and magnesium.
Nitrogen fixation also has been shown to occur in
dead boles (Roskowski 1977, Larsen et al. 1978).
The rates are low but the enormous quantities of
dead wood make the net input to the system
substantial. The effect these nitrogen-fixing,
nutrient-accumulating boles may have on the forest
floor is largely unknown but could be considerable
as the boles frequently occupy 10 percent of the
floor area.
Because boles accumulate nutrients and their
moisture and temperature regimes are more stable
than those of the soil, they are often sites for
germination and subsequent tree development
(Berntsen 1960, Minore 1972). Certain species
particularly prefer down wood as a substrate.
These nurse logs, as they are commonly called,
play an important role in determining the spatial
distribution and species composition of trees in
the forest.
The effect of large woody debris is not limited to
the vegetative nutrient sphere of the ecosystem.
The presence of boles profoundly affects the
insect, bird, small mammal, and ungulate popula-
tions of the forest. Foresters have long known
that survival of many of the so-called forest
insect pests depends on the presence of dead wood
in varying stages of decay. Members of the
Cerambycidae, Buprestidae and Sesiidae are all
dependent on the availability of dead wood, as are
many members of the Hymenoptera. Although they
may damage timber, these insects also serve to
break down and recycle the carbon and nutrients in
the wood and in some cases to rid the forest of
suppressed or unhealthy trees.
Small mammals use the logs as runways and for
protection. These mammals may be critical for the
transport of mycorrhizal spores essential for tree
development (Maser et al. 1978) . The movement of
ungulates is often controlled by impassable down
boles. Thus the presence of down boles may deter-
mine where they feed and travel which in turn may
modify their grazing pressure. In forests such as
the rainforest of the Olympic Peninsula where
grazing pressure by elk is intense and may control
the vegetative structure, boles which impede elk
travel have the potential of determining forest
structure.
Birds are particularly affected by the presence of
both dead standing and down boles. Many species
will nest only in snags and some only in very
particular types of snags (McClellan et al. 1979).
The snags also are used as a source of grubs,
beetles, and other insects. Down boles rarely
provide nesting sites, but the grubs, ants, and
beetles which inhabit them are an important source
of food to birds and mammals alike.
Down logs and snags exert a profound influence on
the ecosystem function of a forest. For this
reason, I chose to study the quality, quantity,
and longevity of down boles and snags as part of a
larger project on the ecosystem function of
streams and forests in the valley bottom of the
South Fork Hon River on the west side of the
Olympic National Park. My objectives were: (1)
to determine the quantity and quality of dead wood
in the mature valley forests and (2) to measure
decay rates of logs of the two dominant tree
species — Sitka spruce (Picea sitchensis (Bong.)
Carr. ) and western hemlock (Tsuga heterophylla
(Raf.) Sarg.).
METHODS
Study Area
The study area was located 5 km inside the Olympic
National Park along the South Fork of the Hoh
River. The valley bottom of the river is quite
flat, fairly broad — often a mile or more across —
and flanked by steep ridges. On these flat, wet
bottoms a series of terraces with associated
forests has developed (see Swanson and Lienkaemper
in this report) . Three general terraces can be
described sequentially away from the river.
Adjacent to the river are narrow bands of red
alder (Alnus rubra Bong.) flats. Next, on what
will be called the lower terrace in this paper, is
an open, grassy, park-like forest of massive Sitka
spruce with occasional vine maples (Acer
circinatum Pursh) and hemlocks. Between this
terrace and the sidewall of the valley, lies the
third or upper terrace. The forest occupying this
terrace has a closed canopy, a moss and fern
floor, and little to no vine maple. Hemlock is
more important on this forest than on the lower
terrace. Because the upper and lower terraces
occupy most of the valley, I chose to study these
rather than the alder flats. A more detailed
description of the forests along the South Fork on
the Hoh is found in McKee et al. in this report.
Field Work
After examining the logs on the forest floor, I
classified them into five types analogous to decay
types developed by Fogel and Cromack (1977) for
Cascade Range Douglas-fir (Pseudotsuga menzlesii
(Mirb.) Franco), but modified for the forest
species and conditions found on the South Fork Hoh
River. Type 1 boles were those with no decay,
fine twigs remaining, and complete bark coverage.
When sliced, the sap and heart wood were clear.
This was the only class that was moss free. Type
2 boles were slightly decayed with most of the
bark present but no fine twigs. When sliced, the
sapwood was rotted but the heart wood sound. Type
3 boles were moderately decayed with some bark
16
present but only stubs of branches remaining.
When sliced, both the sapwood and the heartwood
showed signs of rot; but the bole could still
support itself. Type 4 differed from Type 3 boles
in that Type 4 boles could no longer support them-
selves. Frequently, all the bark was gone. When
sliced, the sapwood was often absent and the
heartwood was a deep red brown and would crumble
into chunks. Type 5 logs were boles noticeable
only by their moss outline on the forest floor.
Bark was entirely absent, and the shape of the
bole was no longer round but oval. When sliced,
the wood was like red powder with little discern-
ible structure or sign of rings. Often a bole
would be a different decay type at opposite ends,
in which case I chose the decay class occupying
the most volume in the log.
In conjunction with other researchers (see
Franklin in this report), two adjacent hectare
plots were laid out on the upper and lower
terraces, respectively. All dead boles greater
than 10 cm in diameter were mapped and their
species and decay class recorded on these plots.
Further description of these hectare plots is
found in McKee et al. in this report. Using these
large maps, I counted the number of wind throws on
the hectares by the number of rootwads present and
the number of windbreaks by the number of snags
with a long intact bole radiating from them.
Trees that died as snags were those snags with no
apparent associated down log.
Four subplots were selected in each hectare plot
for detailed analysis of their dead boles. The
plots were chosen randomly with the stipulation
that no more than two of the subplots could be
contiguous. In each subplot, the length and end
diameters of the down logs were measured, the
species and decay class determined, the coverage
of bark and moss estimated, and the activity of
birds noted. Only bole portions within the plot
were considered. Snags were measured for height
and diameter; characteristics such as bark
coverage, crown presence and bird activity were
noted. On the basis of these characteristics,
snags were later assigned to five decay classes
analogous to those of the boles.
Three samples of wood and bark were taken from'
each down log, taking care to get proportional
samples of sapwood, heartwood, and bark. These
samples were either entire narrow cylinders if the
bole was small or portions of a cylinder if the
bole was large. The samples were cut 4 m from
either end of the bole and at the center. The
sample volume was estimated in the field using
similar geometric forms if it were likely that the
sample would crumble in the process of returning
it to the laboratory. At each sampling point, the
decay state, amount of bark, amount of moss, type
of rot, and diameter of the log were noted.
Lab Work
The samples were taken to a laboratory in Corval-
lis, Oregon, and weighed. Samples whose volumes
had not been computed in the field were either
measured and their volume determined assuming a
geometric form or they were placed in plastic
bags, immersed in water, and their volume deter-
mined by water displacement.
After determining volume, wet weights were
measured. The samples were dried at 60°C for 4
weeks prior to dry weight determinations. Using
the dry weight and volume of each sample, I cal-
culated the density of the wood. The means of the
density of each species and decay class were cal-
culated from these samples. Mean densities were
multiplied by volume measurements of appropriate
decay classes for each of the terraces in order to
calculate the biomass of dead boles on a per hec-
tare basis. Snag biomass was likewise calculated.
To determine the linear and exponential decay
rates of Sitka spruce and western hemlock boles,
the density or logarithm of the density of the
sample was regressed against the sample's age as
determined by the scar. This was done by species
for (1) all the samples, (2) all the samples with
large diameters, and (3) all the samples with small
diameters. The break between large and small was
60 cm for spruce and 30 cm for hemlock. The slope
of the regression gave the decay rate for either
the linear or exponential model. In all cases the
r^ was calculated.
From these data, I calculated the number, volume,
and surface area of down logs and snags by decay
classes and tree species in each of the plots.
Data were then pooled by terrace level.
The segment of the study concerning decay rates
was conducted in areas adjacent to the hectare
plots; 12 Sitka spruce boles and 11 western
hemlock boles were selected. Each bole was a
windthrow or windbreak that had scarred a live
tree when it fell, recording its death date in the
scar. The number of rings laid down since the
scar was determined by cutting a small wedge. The
23 boles covered the range of decay classes except
for Type 5. I had selected five Type 5 boles
which had trees growing on them that I could date
and thus could get a minimum estimate of the bole
age. Upon later microscopic examination, each of
these boles turned out to be a Douglas-fir, a
species no longer of any significance in these
forests.
RESULTS
Maps of two of the four hectare plots are in
Franklin in this report. These maps, which were
made in September 1978, were used successfully the
next year to identify subsequent tree mortality.
Continuing annual surveys are planned.
From these maps, I was able to determine the per-
cent mortality due to windthrow or windbreak.
This varied between upper and lower terrace and
between tree species (table 1). As expected,
windthrow or break was more frequent than snag
formation in the open lower terrace forest. On
the closed canopy upper terrace, wind related
death and snag formation were about equal. Hem-
lock boles were unlikely to become snags on either
terrace although the upper terrace data are ambig-
uous due to the large number of unidentifiable
logs and snags. On either terrace, Sitka spruce
was more likely to become a snag than was hemlock.
17
Table 1 — Numbers of trees that appeared to have died
standing (snags) and been wind killed (windthrow/
windbreak), by terrace level and species; area sam-
pled was 2 ba on each terrace
Upper
terrace
Lower
terrace
Wind throw/
Windthrow/
Species
Snag
windbreak
Snag
windbreak
Sitka spruce
31
21
11
14
Western hemlock
14
33
3
11
Unknown
40
40
1
2
Total
85
94
15
27
The biomass of snags and down wood was markedly
different between the two terraces as was the
relative importance of the two species (fig. 1).
On both terraces, snags represented 20 to 25
percent of the total dead bole biomass. Hemlock
and spruce were of similar importance on the upper
terrace while spruce was more important on the
lower terrace.
Because of the large volume of a single spruce log
on the lower terrace, the surface area occupied by
spruce was low in comparison to its mass (fig. 2).
Thus the lower terrace, which had three-fourths of
the upper terrace's dead wood mass, had only half
its surface area. Eleven percent of the forest
floor was covered by dead wood on the upper terrace
and six percent on the lower. On both terraces,
snags occupied a negligible area.
The decay rates of spruce and hemlock are shown in
table 2 and figures 3 and 4. In addition to the
traditional exponential model, a linear model was
also tried as previous investigators have found
that often the linear model describes bole decay
as well as the exponential model (Grier 1978,
Lambert et al. 1981) . None of the models fit well
due to the highly variable data, but some trends
do emerge. For both species, bole wood taken from
large diameter logs had decayed more slowly than
that taken from small diameter logs. On the
average, hemlock decayed more quickly than spruce,
though this may have been due only to its smaller
size. The linear model, using all the samples,
predicts that a 54-cm spruce bole would be totally
decayed in 141 years and a 37-cm hemlock in 130
years. The exponential model predicts 95 percent
disappearance of spruce in 280 years and hemlock
in 241 years.
ALL DEAD WOOD BIOMASS
UPPER TERRACE
LOWER TERRACE
_ Tsuga
Picea
1 = 70 mt/h»
1=74 mt/ha
jra
n
1=165 mt/ha
Total
Pices
1=72 mt/ha
ws;js\
Tsuga
1 = 44 mt/ha
Total
1 = 122 mt/ha
II III IV V
DECAY CtASS
I II III IV V
DECAY CLASS
Figure 1. — The total biomass of dead wood. The
vertical hatching represents snag biomass.
Figure 2. — The total surface area of the forest
floor occupied by dead wood. The vertical hatching
represents snag area.
TOTAL SURFACE ACRE OCCUPIED BY DEAD WOOD
UPPER TERRACE
LOWER TERRACE
400
300
200
100
0
400
300
200
400
300
200
100
0
Picea 1 = 412 m2/ha
Tsuga
1=525 mVhe
Total
1 = 1.145 mVha
Picea 1=336 m2/h«
Tsuga
1 = 247 mVha
Total
1=611 mVhe
rrrnrn.
I II III IV v
DECAY CLASS
I II III IV V
DECAY CLASS
18
Table 2 — The decay models for Sitka spruce and western hemlock; x = number of
years since death and y = density
Logs > 30 cm (n = 18)
y = -,00251x + .373
(r2 = .150)
y = .375e--0122x
(r2 = .160)
Logs > 60 cm (n = 16)
y = -.00201x + .321
(r2 = .345)
y = .310e-'00881x
(r2 = .269)
WESTERN HEMLOCK
Logs < 30 cm (n = 10)
,00690x + .399
(r<
.878)
-.0177x
y = .368e
(r2 = .609)
SITKA SPRUCE
Logs < 60 cm (n = 19)
y = -.00284x + .378
(r2 = .464)
y = .383 e
-.0119x
(r<
.534)
All logs (n = 28)
1 = -.00310x + .374
(r2 = .227)
y = .363e_-0124x
(r2 = .203)
All logs (n = 35)
! = -.00251x + .356
(r2 = .421)
y = .354e
•.0107x
(r^ = .415)
0.50
0.40ir
E
3
0.30
to
z
0.20
0.10 -
WESTERN HEMLOCK
10
00278*
190)
20 30
YEARS SINCE DEATH
40
50
Figure 3. — The density of
western hemlock plotted
against its age; triangles are
large boles (>30 cm) and cir-
cles are small boles.
0.50 r
0.40
g 0.30
£ 0.20
CO
z
0.10
SITKA SPRUCE
Picea revised October 1979
▲
•
_•
•
•
•
■""^-^
•
y = - 00251x + 356
A
•
^^
(r'=.421)
•
__A
A
•
A
•
A A y = .354e <"°"
(r2=415)
!
- - A
1 1
i i i
i
A ^" —- ^
A
I I I I
10
20
30
40 50 60
YEARS SINCE DEATH
70
80
90
100
Figure 4. — The density of
Sitka spruce plotted against
its age; triangles are large
boles (>60 cm) and circles
are small boles.
19
DISCUSSION
The biomass of dead wood on both terraces was low
but within the 118 to 251 tonnes/ha found in a
chronosequence of 10 mid-elevation forests in the
Cascade Range (Franklin and Waring 1980). Grier
(1978) found 211 tonnes/ha of fallen boles in a
140-year-old Sitka spruce-western hemlock stand at
Cascade Head, Oregon. His value is much higher
than mine, though the forest types are superfi-
cially similar. The Olympic valley bottom spruce-
hemlock forests have low densities relative to
other coastal forests (see McK.ee et al. in this
report) .
Dead bole biomass in excess of 100 tonnes/ha seems
representative of coastal and Cascadian forests of
the Pacific Northwest. For contrast, in the North-
eastern United States, second-growth deciduous
forest has only 28 tonnes/ha (Aber et al. 1978)
and virgin subalpine coniferous forest has 35 to
70 tonnes/ha (Lambert et al. 1981).
The area dead wood occupies (6 and 11 percent) is
very significant in the Hoh valley forests because
regeneration occurs only on dead boles and root
wads (see McKee et al. in this report). These Hoh
forests seem to be an extreme example of Sitka
spruce and western hemlock's preference for
organic substrate (Minore 1972, Berntsen 1960) .
The areal coverage of the forest floor by wood in
either terrace is much lower than the 16 percent
mean areal coverage for the 10 previously cited
Cascadian forests. This is probably due to the
immense Sitka spruce boles present in the Hoh
forests, the low numbers of live stems in the
upper and lower terrace (143 ha-^ and 64 ha~ ,
respectively), and the more rapid decay rates of
hemlock and spruce relative to Douglas-fir (see
McKee et al. in this report). Although snags
occupy only about 30 m2/ha, they are among the
few sites for successful hemlock regeneration as
their height offers protection from elk grazing.
Measuring bole decay rates is an inaccurate pro-
cess at best. Other authors have found low
correlations with linear and exponential decay
models, probably because bole decay is so variable
(Grier 1978, Lambert et al. 1981, Means, personal
communication). One end of a bole may be red
pulp, the other merchantable saw timber. Within a
single disc, I found essentially sound wood next
to fluffy cellulose which, in turn, was adjacent
to cubical brown rot. To minimize this problem, I
took 1 to 9 liters of wood for each sample and
obtained proportionate amounts of the various types
of wood. As the bole diameter becomes larger, the
problem of variable decay becomes increasingly
significant. This is reflected in the lower r2
values for the decay rates of large hemlock and
spruce.
Determining decay rates by the change in density
is conservative as it assumes no change in the
original volume. In reality, sapwood may decom-
pose to the point that it disappears or sloughs
off, and heart wood may compact as it loses its
structure. Either case results in an exaggerated
density value and thus a slower decay rate. In
this study, none of the samples was decayed to the
point of compaction, but some had lost their sap-
wood. Thus my rates are probably slightly low.
BOLE DECA Y
30 40 50
YEARS SINCE DEATH
Figure 5. — The decay rates of boles in several
forests. Value for Cascade Douglas-fir is from:
MacMillan, P. C, J. E. Means, K. Cromack, and
G. M. Hawk. 1979. Douglas-fir decomposition,
biomass, and nutrient capital in the Western
Cascades, Oregon. Unpublished manuscript on file
at Forest Research Laboratory, Oregon State
University, Corvallis, Oregon.
Very few decay rates for boles have been published.
Figure 5 is a compilation of rates found in the
literature compared with those of the Hoh forests.
The Hoh boles decayed more quickly than the
Douglas-fir boles but more slowly than the balsam
fir or tropical rainforest boles. Grier's (1978)
value for Oregon coastal hemlock is quite similar
to the Hoh values for larger hemlock. The forests
of the west side of the Olympics have often been
called temperate rainforests (Franklin and Dyrness
1973), yet they differ in bole decay rates from
those in a true tropical rainforest (Lang and
Knight 1979), a difference due primarily to the
presence of termites in the tropical rainforest.
In the Olympics, insects are a minor component in
the process of bole decay. Insects are also
inconsequential in the subalpine forest, yet the
decay rate is rapid, because of the small size of
the boles ( 15-cm diameter). In fact, the rate
represents the weight loss of all dead boles
standing and down. Standing boles appear to decay
more quickly than down boles in moist forests such
as the Olympic or Pacific Northwest coastal for-
ests (Cline et al. 1980). The Cascadian forests
are drier than the Hoh forests and Douglas-fir is
denser and more resistant to decay (Boyce 1923).
Therefore, although the large spruce boles can
match Douglas-fir for size, they decay more
quickly.
Wood decay, though long a topic of wood products
scientists, has only just begun to show up in the
forest ecology studies. In the Hoh valley forest
where decay occurs quickly, boles represent a huge
pool of biomass, and provide the only site for
regeneration despite their small area of occupancy.
It is essential that we come to understand and
appreciate their interactions with the entire
ecosystem.
20
ACKNOWLEDGMENTS
I wish to thank Dr. Kerrait Cromack for his cheer-
ful advice and help in setting up this study, Ted
Thomas for being my able sawyer and field assist-
ant, and Joe Means for providing me his log com-
puter programs and even more for running them.
LITERATURE CITED
Aber, J. D. , D. B. Botkin, and J. M. Milillo.
1978. Predicting the effects of different har-
vesting regimes on forest floor dynamics in
northern hardwood. Can. J. For. Res. 8:306-315.
Berntsen, C.
1960. Planting Sitka spruce and Douglas-fir on
decayed wood in coastal Oregon. USDA For. Serv.
Res. Note PNW-197, 5 p. Pac. Northwest For. and
Range Exp. Stn. , Portland, Oreg.
Grier, C. C.
1978. A Tsuga heterophylla-Picea sitchensis
ecosystem of coastal Oregon: decomposition and
nutrient balances of fallen logs. Can. J. For.
Res. 8:198-206.
Lambert, Robin L., Gerald E. Lang, and William A.
Reiners.
1981. Loss of mass and chemical change in
decaying boles of a subalpine balsam fir
forest. Ecology 61(6) :1460-1473.
Lang, G. E., and D. K. Knight.
1979. Decay rates for the boles of tropical
trees. Biotropica 11(4) :316-317.
J., M. F. Jurgensen, and A. E. Harvey.
2 fixation associated with wood
Larsen, M
1978. N
decayed by some common fungi in western Mon-
tana. Can. J. For. Res. 8:341-345.
Boyce, J. S.
1923. Deterioration of windthrown timber on the
Olympic Peninsula, Washington. U.S. Department
of Agriculture Tech. Bull. No. 104. 28 p.
Maser, C. , J. Trappe, and D. Ure.
1978. Fungal-small mammal interrelationships
with emphasis on Oregon coniferous forests.
Ecology 59(4): 799-809.
Cline, S. P., A. B. Berg, and H. M. Wight.
1980. Snag characteristics and dynamics in
Douglas-fir forests, western Oregon. J. Wildl.
Manage. 44:773-786.
Fogel, R., and K. Cromack.
1977. Effect of habitat and substrate quality
on Douglas-fir litter decomposition in western
Oregon. Can. J. Bot. 55:1632-1640.
Franklin, Jerry F. , and C. T. Dyrness.
1973. Natural vegetation of Oregon and Washing-
ton. USDA For. Serv. Gen. Tech. Rep. PNW-8, 417
p., illus. Pac. Northwest For. and Range Exp.
Stn., Portland, Oreg.
Franklin, J. F. , and R. H. Waring.
1980. Distinctive features of the northwestern
coniferous forest: development, structure and
function. In Richard Waring, ed. Proceedings,
20th Annual Biology Colloquium. Oreg. State
Univ. Press. Corvallis, Oreg.
McClellan, B. R. , S. S. Frissell, W. C. Fischier,
and C. H. Halvorson.
1979. Habitat management for hole-nesting birds
in forests of western larch and Douglas-fir. J.
For. 77:480-484.
Mi no re, D.
1972. Germination and early survival of coastal
tree species on organic seed beds. USDA For.
Serv. Res. Pap. PNW-135, 6 p. Pac. Northwest
For. and Range Exp. Stn., Portland, Oreg.
O'Neill, R. V., W. F. Harris, B. S. Ausmus, and
D. E. Reichle.
1975. A theoretical basis for ecosystem
analysis with particular reference in element
cycling. Pages 28-40 in Mineral Cycling in
Southeastern Ecosystems. F. G. Howell, J. B.
Gentry and H. H. Smith, eds. ERDA Symposium
Series (CONF-74013) .
Roskowski, J. P.
1977. Nitrogen fixation in northern hardwood
forests. Ph. D. thesis, Yale Univ. New Haven,
Conn. 112 p.
21
Structure, Composition, and Reproductive Behavior of Terrace Forests,
South Fork Hoh River, Olympic National Park
Arthur McKee, George LaRoi, and Jerry F. Franklin
ABSTRACT
Mature forests of Plcea sitchensis and Tsuga heterophylla varied with terrace level. Upper terraces had
denser stands, greater numbers of Tsuga, and understories of Vaccinium, ferns, and mosses. Lower terraces
had open stands with understories of Acer circinatum and grasses. Tree reproduction occurred primarily on
down logs. Less than 1 percent occurred on ground humus. Picea reproduction numbers and survival rates
were superior to Tsuga. Tsuga reproduction may have exceeded that of Picea earlier. Both similarities
and differences exist with Fonda's Hoh River model. Picea was apparently climax in these terrace forests
in contrast to other coastal types.
Arthur McKee, Department of Forest Science, Oregon
State University, Corvallis, Oregon 97331; George
LaRoi, Department of Botany, University of
Alberta, Edmonton, Alberta, Canada T6G 2E9; and
Jerry F. Franklin, U.S. Department of Agriculture,
Forest Service, Pacific Northwest Forest and Range
Experiment Station, Forestry Sciences Laboratory,
3200 Jefferson Way, Corvallis, Oregon 97331.
22
Franklin (in this report) has described the rela-
tionships of the vegetation studies to the other
portions of the South Fork Project. Plant com-
munity analyses in the South Fork were directed
toward collecting baseline vegetation data in a
manner that would facilitate their interpretation
relative to topographic position and proximity to
aquatic habitats. Previous studies by members of
the research team had shown the importance of woody
debris in mediating erosional processes (Swanson
1980). Information was sought to examine these
interactions in a coastal ecosystem and the results
are presented by Swanson and Lienkaemper in this
report.
From research in the main valley of the Hoh River,
Fonda (1974) proposed a successional model for
river terrace forests that interpreted successively
higher terraces as a series of serai stages tending
toward a climax forest of Tsuga heterophylla. Ter-
race formation occurs by meandering of the river
and periodic flooding. The wide valley floor of
the South Fork of the Hoh River with its extensive
terraces seemed an appropriate place to test the
generality of Fonda's (1974) model.
More knowledge of the successional roles of Picea
sitchensis and Tsuga heterophylla was needed to
further evaluate the above model. In mature
forests of the coastal fog belt (Franklin and
Dyrness 1973) and on alluvial terraces bordering
the lower reaches of coastal rivers (Cordes 1972),
the two species compete strongly for dominance.
In some habitats P. sitchensis seems unable to
co-exist indefinitely with T. heterophylla in the
absence of disturbances because it is less shade
tolerant (Fowells 1965). P. sitchensis succeeds
Alnus rubra in alluvial terraces for some distance
inland, becoming the dominant tree species (Cordes
1972, Fonda 1974). It may be more tolerant of sea-
sonal flooding in these sites than T. heterophylla.
As one of the major attractions of the Olympic
National Park, the stability and successional fate
of the terrace forests is of extreme interest.
Therefore, regeneration behavior of P_. sitchensis
and T_. heterophylla was a major part of the vege-
tation studies.
This paper discusses the vegetation of the South
Fork terraces in terms of upper versus lower ter-
race stands, but the collation of our terms with
those of the geomorphologists is critical. At
least six terrace surfaces have been recognized in
the study area (see Swanson and Lienkaemper in
this report). Surfaces 1, 2, and 3 range from
fresh gravel bars to low terraces with young A.
rubra stands. Their surfaces 4, 5, and 6 are-
occupied by mature, conifer-dominated forests and
are the locale of the studies reported here. Data
from surfaces 5 and 6 are combined for our "upper
terrace" values. Surface 4 is our "lower terrace."
Although it is difficult to draw exact comparisons
between the adjacent river valleys, our lower
terrace is believed roughly equivalent to Fonda's
(1974) first terraces, and our upper terrace is
roughly equivalent to his second terrace.
METHODS
Two different vegetation sampling procedures were
employed in this study to accommodate the diverse
needs of the group (see Franklin in this report).
Transects allowed us to examine the interactions
between topographic position, vegetation, and
aquatic habitats. A point-quarter sampling method
was used at 50-m intervals on these transects.
Herbaceous cover was estimated in eight microplots
(20- x 50-cm) around each point. Shrub cover was
estimated by line intercept along the transects.
The remainder of the vegetation sampling was con-
ducted on the four large (1-ha) permanent plots
(see Franklin in this report). Detailed sampling
of down logs for dimensions and decay class is
described by Lambert, in this report, who provided
the basic data utilized here on log numbers and
surface area.
Restricted random sampling was performed on each
of the two terrace stands of the immature tree
subpopulations (stem £ 8 m tall). Five major sub-
strate types were sampled: ground duff and humus,
tree stumps and uprooted tree bases ("root wads"),
P. sitchensis logs, T. heterophylla logs, and logs
unidentifiable as to species. The three log sub-
strates were further subdivided by a decay class
scheme (see Lambert in this report). A logarith-
mic height classification system was used in tal-
lying seedlings and saplings in which the height
range of each taller class is doubled: e.g.,
(1) < 0.125 m; (2) 0.125 - 0.25 m; (3) 0.25 - 0.5
m, etc. This was done to achieve a greater reso-
lution in smaller size classes.
Biomass estimates were made from data collected on
the four permanent plots using allometric equations
developed by the Coniferous Forest Biome (Gholz et
al. 1979). Heights of selected trees were measured
on the permanent plots. Ages of overstory domi-
nants were determined by increment cores taken
along the transects.
RESULTS
Structure and Composition
Forests on the valley floor of the South Fork of
the Hoh River were dominated by P. sitchensis and
T. heterophylla except for recently created,
relatively narrow terraces along the main river
channel, which were dominated by A. rubra. Much
of the valley floor of the South Fork consists of
two relatively distinct terraces, however, and
most of the following discussion is a comparison
of these upper and lower terraces. Although there
are similarities, the data from the transects and
permanent plots are evidence that different ter-
race levels were occupied by stands that differ
substantially in structure and composition.
Both upper and lower terraces had similar ages and
heights of the overstory Picea and Tsuga. Mean and
maximum ages at 1.5 m above ground were 220 and 266
years on the upper and 205 and 258 years on the
lower terrace. The tree strata on both terraces
consisted of a tall tree layer of P. sitchensis 75
to 80 m in height and a medium tree layer of T.
heterophylla and P. sitchensis of 45 to 55 m. Max-
imum heights measured on both terraces exceeded
85 m. Mean heights and ages were slightly greater
on the upper terrace, but sampling was insufficient
for a test of significance.
23
Table 1 — Mean densities, basal areas, and diameters
by species for upper and lower terrace forests of
the South Fork of the Foh River
esi'
Densl
ty
Basal
area
Mea
diame
ter
Sped
Upper
Lower
Upper
Lower
Upper
Lower
Number
per
Square
meter
hectare
per hectare
Centimeters
PISI
57.8
33. 1
61.9
52.8
90.4
118.4
TSHE
THPL
79.9
2.1
24.7
2/0.3
15.8
1.0
10.3
1/0.8
45.8
73.0
64.9
2/176
PSME
ALRU
1.7
2/0.3
5.3
2.9
2/0.03
1.9
114.0
1/28.0
64.9
AC HA
ABAM
y o'i
0.7
i'o.l
0.4
2/5l"o
84.0
All ;
pedes
142
64
81.8
66.3
65.6
93.3
I/spe
des are: PISI
= Picea
si tchensis
TSHE =
Ts
uga
heterophylla
THPL -
Thuja pi
icata, PSME = Pseudotsuga menz
lesil,
ALRU - Alnus rubra , ACMA = Acer aacrophyllum, and ABAM = Abies
ama bills.
'Only one Individual in the sample.
1 00 1 50 200
DIAMETER (cm)
Figure 1. — Diameter distributions for all Picea
sitchensis and Tsuga heterophylla sampled which
were over 5 cm in diameter on upper and lower
terraces.
Upper and lower terrace stands differed substan-
tially in several structural and compositional
features (table 1 and fig. 1). Total density and
basal area were greater in upper terrace than
lower terrace stands: 142 vs. 64 stems/ha and
81.8 vs. 66.3 m2/ha. Mean diameters, on the
other hand, were greater on the lower than the
upper terrace. P. sitchensis averaged 118-cm
d.b.h. on the lower and 90-cra d.b.h. on the upper
terrace, while T_. heterophylla averaged 65- and
46-cm d.b.h. respectively. The larger diameters
on the lower terraces may have been due to wider
spacing which results in reduced competition or to
better site conditions or both. Diameter distri-
butions of P^. sitchensis and J_. heterophylla
showed that, except for the smallest size class,
V_. sitchensis had peak densities at substantially
larger size classes than J_. heterophylla on both
terraces (fig. 1). The bimodal nature of the I\
sitchensis diameter distributions on both terraces
should be noted. The implications of this distri-
bution will be discussed later in the section on
regeneration.
The biomass estimates in table 2 reflect some of
the structural variation encountered in upper and
lower terrace stands. Total biomass was measured
in metric tonnes per hectare (t/ha) and averaged
more in the upper terrace permanent plots than in
the lower terrace. Considerable variation existed,
however, in the biomass of P^. sitchensis in the
upper terrace plots and in T_. heterophylla in the
lower terrace plots. Total biomass was quite high
on both terraces for 200- to 250-year-old stands,
indicating the productive nature of the Picea-
Tsuga forests.
Composition differs dramatically between upper and
lower terrace forests. T_. heterophylla was of much
greater importance in upper terrace stands as shown
by the density and basal area values (table 1).
Minor tree species such as Pseudotsuga menziesii
and Abies amabilis seemed confined to upper ter-
races. Alnus rubra was much more important on the
lower terraces (tables 1 and 2).
Understory composition also differed between the
two terraces (table 3). Composition of the shrub
layer shifted from Acer circinatum-dominated to
Vaccinium-dominated. Cover of Acer circinatum
dropped from 28.2 to 2.4 percent as one went from
lower to upper terrace stands, while Vaccinium
cover went from 0.6 to 10.8 percent. Total shrub
cover was twice as great on the lower terrace.
The herbaceous layer of lower terrace stands was
dominated by grasses and forbs with 25.4- and
37.3-percent cover, respectively. As with Acer,
grass cover dropped dramatically in upper terrace
stands. Mosses and ferns became much more impor-
tant in the upper terrace forests. Total cover of
forbs decreased slightly in upper terrace forests.
Species composition of the forbs was quite similar
on both terraces.
24
Table 2 — Biomass estimates (metric tonnes per hectare) for four permanent 1-ha plots
established September 1978 in upper and lower terrace forests of the South Fork of
the Hoh River, Olympic National Park. Calculated for stems > 15-cm d.b.h.
Species
1/
Bole and bark
Upper
Lower
Branches
Upper
Lower
Foliagei/
Upper
Lower
Total
Upper Lower
PISI
TSHE
THFL
PS ME
ALRU
Total
x
sd
442.4
428.4
50.7
51.7
740.5
446.8
86.2
53.4
110.6
36.4
33.6
12.8
62.7
154.6
18.3
56.8
2.6
10.8
0.5
1.9
0.3
—
0.2
—
17.2
—
0.8
—
22.9
—
1.1
—
—
5.0
—
0.6
—
4.1
—
0.5
572.8
482.6
85.6
67.0
829.7
605.4
106.1
110.6
701.2
544.0
95.8
88.8
181.7
86.8
14.5
30.8
4.2
4.0
497.3
484.1
6.9
4.2
833.6
504.3
6.3
2.1
150.4
53.3
3.6
8.1
84.6
219.4
0.2
0.7
3.3
13.4
0.1
—
1.1
—
0.2
—
18.3
—
0.3
—
24.2
—
—
0.03
—
5.6
—
0.02
—
4.6
10.9
6.9
669.3
556.4
11.1
12.3
946.8
728.3
11.0
9.6
808.0
642.4
0.1
3.8
196.2
121.6
i/species are: PISI = Picea sitchensis, TSHE = Tsuga heterophylla, THPL = Thuja
plicata, PSME = Pseudotsuga menziesii , and ALRU = Alnus rubra.
_'Foliar biomass estimates for Picea are conservative because the equations used
(Gholz et al. 1979) are partially based on wind-trimmed trees.
Table 3 — Mean composition of shrub and herbaceous layers of forest communities on
the upper and lower terraces of the South Fork of the Hoh River
Shrub layer!/, % Cover
Herbaceous layer, % Cover
ACCI VAsp RUSP MEFE £ Forbes Grasses Mosses Ferns
Upper terrace 2.4 10.8 1.9 .03 15.1 29.8 4.8 64.3 13.5
Lower terrace 28.2 0.6 1.8 0.0 30.6 37.3 25.4 42.4 6.8
_'Shrub species abbreviations are: ACCI = Acer circinatum, VAsp = Vaccinium
species, RUSP = Rubus spectabilis, and MEFE = Menziesia f erruginea.
25
Table 4--I' i cea si tr hens is and Tsuga heterophylla
sabpopulation densitv (stems < 8 m tall/m2) on
the 5 major substrate types and 4 major log decay
classes in the upper and lower terrace forests,
South Fork lloh River, Olympic National Park
Upper
terrace
Lower
terrace
Substrate type and
log decay class
Picea
Tsuga
Picea
Tsuga
Substrate type:
Picea logs
36.0
15.3
19.3
6.6
Tsuga logs
29.9
9.6
11.3
4.3
Unknown logs
30.0
7.5
11.5
5.1
Stumps and root wads
5.1
1.6
7.7
2.9
Ground humus
0.08
0.01
0.08
0.01
Log decay class:
2. Early
24.8
15.5
8.3
2.0
3. Middle
38.5
10.7
21.7
7.7
4. Late
28.6
9.2
11.9
5.4
5. Very late
28.2
8.2
11.9
4.7
One interesting aspect of species composition
concerned the naturalized Eurasian weeds. These
exotic species were confined almost entirely to
lower terrace stands and recently formed terraces
and gravel bars. The only exotic species encoun-
tered in the upper terraces were Agrostis alba and
Prunella vulgaris, and these were rare. The other
exotic species — Poa trivialis, Ranunculus repens,
Rumex acetosella, Rumex crispus, Trif olium repens,
and Lactuca serriola — seem restricted to the lower
terraces. Agrostis alba, Poa trivialis, and
Ranunculus repens were important components of the
lower terrace herbaceous layer with covers fre-
quently exceeding 20 percent. The other species
tended to be locally important or rare. Fonda's
(1974) data show these Eurasian weeds similarly
distributed on the terraces in the main stem of
the Hoh River. Historically, traffic of both
humans and livestock has been heavier in the more
open lower terraces. These factors and the more
frequent disturbance by flooding on the lower ter-
races were probably the reasons for the current
distribution of exotics.
Forest Tree Reproduction
Regeneration on Different Substrates
The density of seedlings and
8 m tall growing on the five
on the two terraces is given
sitchensis and T. heterophyl
ficulty establishing on grou
terrace stands. Down logs a
able site for establishment,
are better recruitment sites
logs for both tree species,
confirmed by chi-square test
observed by Minore (1972).
saplings less than
major substrate types
in table 4. Both P.
la obviously had dif-
nd humus in the two
re a much more favor-
P. sitchensis logs
than T. heterophylla
This difference is
s and also has been
Total recruitment densities of P_. sitchensis were
higher than those for T_. heterophylla on all five
substrates in both terrace stands. Environmental
conditions during the life span of these seedlings
and saplings have clearly favored P. sitchensis
over T_. heterophylla.
Regeneration on Different Log Decay Classes
Densities of seedlings and saplings on four dif-
ferent log decay classes also are tabulated in
table 4. Both terraces lacked sufficient logs in
class 1 to provide good tree seedling density
estimates. P_. sitchensis density was highest on
class 3 logs on both terraces. The maximum den-
sity of T_. heterophylla regeneration, however, was
on class 2 logs in the upper terrace stands and
class 3 logs in the lower terrace stands.
Chi-square tests support two conclusions: (1) tree
recruitment potentials of fallen logs in the study
area change significantly with decomposition stage
and (2) the off-log environments of the two ter-
races exert a strong influence on the recruitment
potential of different log decay classes. Seed-
ling and sapling density is almost always greater
on the upper than lower terrace for a given sub-
strate type and species.
Total Regeneration
The data on densities on different substrates and
decay classes reveal much about the behavior at
the two tree species but do not show regeneration
in the two terrace stands. Total density per
hectare is the product of the above densities and
the amount of surface area per hectare occupied by
the different substrate types and decay classes.
These calculations show J^. sitchensis regeneration
was approximately three times more abundant than
T. heterophylla in both upper and lower terrace
stands (table 5).
The importance of logs as a recruitment site is
dramatically shown in table 5. Logs provided 96
and 88 percent of the P. sitchensis recruitment on
the upper and lower terraces, respectively. Pro-
portions of T_. heterophylla were 97 and 93 percent.
Class 3 logs are obviously of special importance
(table 5). On a per hectare basis, they supported
the largest number of seedlings and saplings less
than 8 m tall for both species on both terraces.
Survivorship
Total densities (table 5) are interesting but do
not reveal differences between species in abun-
dance of different size classes. The abundance of
P. sitchensis might be largely restricted to the
smallest size classes, T. heterophylla to the
largest, which would have vastly different impli-
cations for successional trends. Abundance data
are presented in figure 2 as height-based survi-
vorship curves for the two species on the two
terraces. The regeneration sampling provided data
for the first seven height classes, and the tran-
sect and permanent plot samples provided data for
larger height classes.
26
Table 5 — Calculated densities (stems £ 8 m tall) of
Picea sitchensis and Tsuga heterophylla (ha-*) on 5
major substrate types and 4 major log decay classes
in the upper and lower terrace forests, South Fork
Hoh River, Olympic National Park
Upper terrace
Lower terrace
Substrate type and
log decay class
Picea
Tsuga
Picea
Tsuga
Substrate type:
Picea logs
14,500
6,170
5,810
1,980
Tsuga logs
15,300
4,910
2,420
930
Unknown logs
4,920
1,230
600
260
Stumps and root wads
550
170
440
160
Ground humus
700
90
750
90
Total
35,970
12,570
10,020
3,420
Log decay class:
2. Early
7,030
4,340
500
116
3. Middle
19,000
5,250
7,530
2,660
4. Late
7,800
2,500
1,730
780
5. Very late
820
170
80
30
The survivorship curves show P_. sitchensis to be
more abundant than T_. heterophylla on both terrace
forests for the first eight height classes (£16 m
tall) (fig. 2). The environmental conditions dur-
ing the recent past have clearly favored P. sit-
chensis recruitment and survival. Such has not
always been the case, however.
The survivorship curves for both terraces show an
increase in T_. heterophylla abundance in the larger
size classes with a concomitant reduction of P.
sitchensis (fig. 2). The diameter distributions
shown in figure 1 revealed a greater abundance of
J_. heterophylla in the intermediate size classes,
sandwiched between the peaks of the bimodal distri-
bution of P. sitchensis. Hence, T. heterophylla
regeneration was favored over P. sitchensis at
some time in the past. The relative position of
the survivorship curves for that period would have
been reversed for both terraces, with T. hetero-
phylla more abundant than P. sitchensis in the
smaller size classes.
100,000
10,000 -
UJ
or
<
H
O
UJ
X
fr
UJ
o_
en
UJ
I-
100,000
10,000
or
<
r-
U
UJ
X
or
UJ
a.
UJ
(f)
10 II
HEIGHT CLASS
HEIGHT CLASS
Figure 2. — Survivorship curves for Picea sitchensis
and Tsuga heterophylla on lower and upper terrace
forests in the South Fork of the Hoh River; each
height class is twice as large as its predecessor
(i.e., 1 = < 0.125 m, 2 = 0.125 to 0.25 m, 3 = 0.25
to 0.5 m, 4 = 0.5 to 1 m, 5 = 1 to 2 m, 6 = 2 to
4m, 7 = 4 to 8 m, 8 = 8 to 16 m, 9 = 16 to 32 ra,
10 = 32 to 64 m, and 11 = > 64 m tall).
27
DISCUSSION
The composition and structure of the terrace for-
ests of the South Fork of the Hoh River have both
similarities and differences with Fonda's (1974)
terrace-based model developed for the main Hoh
River drainage. The higher terraces supported
increasing amounts of J_. heterophylla in both the
South Fork and the main Hoh River valleys. Grasses
were an important component of the lower terraces
in both systems; mosses increased in cover on the
upper terraces. Forb cover remained relatively
constant on all the terraces, and naturalized Eur-
asian weeds were confined almost entirely to the
lower terraces in both valleys.
There were slight differences in the terrace for-
ests between the two valleys, however. Alnus
rubra is present in the lower terrace forests in
the South Fork valley as well as on recent allu-
vium. The lower terrace forests in the South Fork
appeared to be a much older version of Fonda's
(1974) first terrace forest and lacked Populus tri-
chocarpa. The upper terrace stands of the South
Fork appeared to be intermediate to the second and
third terrace forests of Fonda (1974). The upper
and lower terrace stands of the South Fork had
lower densities and higher basal areas than their
analogs in the main Hoh River valley. Shrub cover
was higher in the South Fork terraces with Acer
circinatum much more abundant on the lower terraces
and Vaccinium species more abundant on the upper
terrace.
More important differences concerned the ages of
the terrace forests and the role of P. sitchensis
in the South Fork stands. In contrast to Fonda's
(1974) model, the forests of the upper and lower
terraces had dominants of about the same age and
thus cannot be viewed as serai stages in a se-
quence of forest development. As shown by the
survivorship curves, P. sitchensis regeneration
was currently favored over T. heterophylla on both
terraces. This is in contrast to what Fonda (1974)
reports for even his third terrace forests which
average two-thirds the basal area of our upper ter-
race stands, and thus should favor Picea by virtue
of being less shaded.
Franklin and Dyrness (1973) suggest a climax role
for P. sitchensis on alluvial habitats, in contrast
to its serai role throughout most of the coastal
zone. Fonda (1974), on the other hand, indicates
J_. heterophylla is the climax species on the older
terraces. No evidence exists in our data for a
directional change from P. sitchensis to T_. heter-
ophylla dominance in the South Fork valley. P.
sitchensis was currently replacing itself on both
terraces and should maintain, if not actually in-
crease, in importance relative to T_. heterophylla.
A successional shift to Tsuga dominance is not
apparent. This is consistent with Cordes' (1972)
findings in valley Picea-Tsuga stands in British
Columbia.
Several factors could be responsible for the per-
sistence of Plcea in these alluvial forests. Many
of the stands were park-like with widely spaced
stems and numerous openings ranging up to several
hectares in size. Such conditions would be more
favorable for Picea which is less shade tolerant
than Tsuga. An often-cited factor is grazing by
Roosevelt elk which are believed to feed selec-
tively on the Tsuga. The importance of browsing
has yet to be demonstrated, however, and could not
account entirely for the better survival of Picea
reproduction. Picea currently demonstrated supe-
rior survival from the smallest size classes,
presumably below the size of material typically
taken by elk.
It is important to note, however, that within the
ages of the stands in the South Fork, T. hetero-
phylla regeneration has been favored relative to
that of P. sitchensis. The survivorship curves
for both upper and lower terraces show a curious
reversal in the abundances of pole-sized individ-
uals of the two species. The reasons for this
oscillation in regeneration of the two species is
not clear. It could be related to climatic fluc-
tuation or to changes in the population of Roose-
velt elk. Elk populations were low early in this
century, which might account for the wave of
larger-sized T. heterophylla. Age data collected
for Tsuga do not support this hypothesis, however
(i.e., trees sampled did not originate uniformly
in the period 1890 to 1910). Fluctuations of
either climate or elk herds could create unstable
size structures with oscillations in abundance due
to time-lag effects.
The lack of correlation between Fonda's (1974)
model and the terrace forests of the South Fork is
not surprising. Each terrace valley has had its
own history of disturbances since the retreat of
the glaciers from the valleys. Flooding patterns
have not been identical. Moreover, the South Fork
valley is distinctive in the breadth of the valley
floor relative to the total width of the drainage.
Colluvial and alluvial depositions from the valley
walls and tributaries, which might alter the basic
patterns in these terrace forests, are much less
important in the South Fork than in the main Hoh
River valley. For example, Acer macrophyllum
groves which were confined to colluvial fans (Fonda
1974) were almost absent in the South Fork. The
fire history of the two valleys could also be very
different, creating an array of serai stand condi-
tions that correlate poorly.
Results of this study clearly show the importance
of logs and decaying wood for forest perpetuation.
Recruitment would be sparse, indeed, but for these
substrates. The superiority of P_. sitchensis logs
as a regeneration site remains to be explained.
The factors which severely limit regeneration on
the forest floor are doubtless many, including com-
peting vegetation and possibly disease. The conse-
quences are clear, however, from the extremely rare
seedlings not associated with down logs, stumps, or
root mounds. Removal of these materials from ter-
race stands would clearly limit regeneration.
28
Implications for lands managed for timber produc-
tion in the vicinity of the Park are clear. Rot-
ten wood is an important substrate for seedling
establishment on cutovers, especially where shrub
competition is severe. Rotten logs, stubs, and
root wads which are potential sites for seedling
establishment should be viewed as assets in regen-
eration problem areas on the west side of the Olym-
pic Peninsula.
CONCLUSIONS
The mature forests described on terraces in the
South Fork of the Hoh River provide an outstanding
sample of the valley-bottom stands sometimes refer-
red to as Olympic rainforest. Details of composi-
tion and structure do vary in other parts of the
South Fork and in the main river drainages of the
western Olympic Mountains — main Hoh, Quinault,
Queets, and Bogachiel Rivers. Nevertheless, some
general conclusions about the valley-bottom for-
ests are possible. Landform is, as indicated by
Fonda (1974), of major importance in forest compo-
sition and structure. The mature Picea-Tsuga
forests are often relatively open. Logs are of
overwhelming importance for recruitment of new
conifers; nurse logs are an essential structure
rather than an interesting novelty. A variety of
factors appears to make Picea as much a potential
climax species as Tsuga. Roosevelt elk intuitively
appear important influences on vegetative composi-
tion and dynamics, but quantitative data still are
absent.
LITERATURE CITED
Cordes, Lawrence D.
1972. An ecological study of the Sitka spruce
forest on the west coast of Vancouver Island.
Ph.D. thesis, Univ. of British Columbia, Van-
couver, Canada. 452 p., illus.
Fonda , R. '4 .
1974. Forest succession in relation to river
terrace development in Olympic National Park,
Washington. Ecology 55(5) :927-942, illus.
Fowells, H. A.
1965. Silvics of forest trees of the United
States. USDA Handb. 271. 262 p. , illus.
Franklin, Jerry F. , and C. T. Dyrness.
1973. Natural vegetation of Oregon and Washing-
ton. USDA For. Serv. Gen. Tech. Rep. PNW-8,
417 p., illus. Pac. Northwest For. and Range
Exp. Stn. , Portland, Oregon.
Gholz, H. L., C. C. Grier, A. G. Campbell, and
A. T. Brown.
1979. Equations for estimating biomass and leaf
area of plants in the Pacific Northwest. Oreg.
State Univ., For. Res. Lab. Res. Pap. 41, 37 p.
Minore, Don.
1972. Germination and early growth of coastal
tree species on organic seed beds. USDA For.
Serv. Res. Pap. PNW-135, 18 p., illus. Pac.
Northwest For. and Range Exp. Stn., Portland,
Oregon.
Swanson, Frederick J.
1980. Erosion as an ecosystem process. In
Richard Waring, ed . Proceedings 20th Annual
Biology Colloquium. Oreg. State Univ. Press.
Corvallis, Oregon. In Press.
29
Interactions Among Fluvial Processes, Forest Vegetation, and Aquatic Ecosystems,
South Fork Hoh River, Olympic National Park
Frederick J. Swanson and George W. Lienkaemper
ABSTRACT
Interactions among fluvial processes and forest vegetation created a variety of landforms, plant communi-
ties, and aquatic habitats in the South Fork Hoh River. We distinguished six geomorphic surfaces based on
differences in vegetation and elevation relative to low water level. Relations between high flows and
forest vegetation vary from one surface to another. Flood effects included inundation, bank cutting, sur-
face scour, deposition, and transport of large organic matter. Geomorphic processes have created four
distinctive aquatic habitats in the valley: main river channel, off-channel areas along the main stem,
and valley-wall and valley-floor tributary streams.
Frederick J. Swanson and George W. Lienkaemper,
U.S. Department of Agriculture, Forest Service,
Pacific Northwest Forest and Range Experiment
Station, Forestry Sciences Laboratory, Corvallis,
Oregon.
30
\^ ftPPROX TRUE H
J&JP^EHttf " v& ' ~
n^" ^"^w^m
■ /t:-x
^.P*m2--^
^s^^^%^.^^
y^4
^^O^^^^^^i^'^^'^ j^W^' rf^ t
s-v — v\V\ A \=i)
^V\ ><.
■^22_£! ^^'iL^^2^-
j^s^^J ^W \ v/
^Cj$^
S ^-**R A /
Ksg^* r^
| [ SUPFACE 1 - Grovel bor, unveoetoted, 0-1 m ALW
;-----\^^-~Oc^^N^S
^£i^^^^^^
| ] SURFACE 2 - Alder t>~25yr. I-I8m ALW
\ J SURFACE 3- Alder onrj spruce to 100 yr . !5-2mALW
f-:.;.] SURFACE 4 - Spruce 200- yr, 2-3m ALW
■
f£2 SURFACE 5 - Spruce 200 » yr. ~ 5m ALW
^1 MAIN STEM, flowing 9/76
S SLUMP BLOCK
■■ SURFACE 6 - Spruce 200 • yr , ~8m
["'IhIGH FLOW CHANNEL, dry 9/78
^ LARGE ORGANIC DEBRIS
ALW ABOVE LOW WATER 9/78
/ TRIBUTARY
.^^ FLOATED ORGANIC DEBRIS
Figure 1. — Geomorphic
surfaces, channel posi-
tion, and large organic
debris in a section of
South Fork Hoh River.
Mapped by pace and
compass by G.W.
Lienkaemper .
INTRODUCTION
Landforms and geomorphic processes are important
factors in development of most terrestrial and
aquatic ecosystems. Interactions between physical
and biological features and processes are espe-
cially well developed along glacier-fed rivers
flowing through heavily forested, glacially carved
valleys such as the Hoh River. On the broad valley
floor of the Hoh River system, fluvial geomorphic
processes create landforms providing sites for ter-
restrial and aquatic ecosystem development that
contrast markedly with valley-wall sites. Fluvial
processes regulate the development of these ecosys-
tems in areas subject to flooding and sedimentation.
In the geomorphology phase of the interdisciplinary
South Fork Hoh River study, we examined these rela-
tionships by addressing three specific objectives:
1. Map valley-floor geomorphic surfaces and chan-
nel features and age associated trees.
2. Examine relations among fluvial processes and
vegetation along the main river channel.
3. Define relations among fluvial processes and
vegetation along the main river channel.
We briefly report results of each of these study
phases to provide the physical environmental con-
text within which to view the study results.
FLUVIAL SURFACES AND CHANNELS OF THE MAIN STEM
At low and moderate flow conditions, the main stem
of the South Fork Hoh River meanders within its
broad, gravel-floored, flood channel. Within the
study reach, about 2.5 km east of the western
Olympic National Park boundary, the unvegetated
flood channel is about 100 m wide, less than 10
percent of average valley-floor width. The steep
(60-percent) valley side slopes, carved by gla-
ciers as recently as the latest Wisconsin advance
(Crandell 1965, Heusser 1974), end abruptly at the
valley floor. The valley bottom is partially
filled with younger outwash gravels that form a
complex set of terraces.
Figure 2. — Surveyed cross section of valley bottom
along transect A-A' (see fig. 1).
The 1 000-m study reach contained six geomorphic
surfaces distinguishable on the basis of vegetation
and elevation above river level (figs. 1 and 2).
Successively higher surfaces up to Surface 5 bore
forest communities in progressively more advanced
stages of development and greater amounts of large
woody debris produced by forests on the surface.
Emergent, unvegetated gravel bars of Surface 1
extended to 1 m above the late summer river level
at the time of the study (mid-September 1978).
Surface 2 had alder, Alnus rubra, thickets of trees
up to 30 years old. Alder and spruce, Picea sit-
chensis, trees on Surface 3 were up to more than
100 years in age. Old-growth, open-grown spruce
and western hemlock, Tsuga heterophylla, on Sur-
face 4 ranged in age up to 258 years, based on a
sample of about 20 trees (see McKee et al. in this
report). Individual spruce and hemlock trees on
Surface 5 did not appear to be significantly older.
McKee et al. (in this report) observed a maximum
age of 266 years in a sample of about 20 trees.
Unlike Surface 4, however, many large, down boles
in advanced stages of decay littered the forest
floor of Surface 5, suggesting that the community
might have been much older than the oldest living
individuals. This hypothesis was favored by the
occasional occurrence of well-rotted Douglas-fir
logs on Surfaces 5 and 6 (see Graham in this re-
port). An alternative hypothesis is that greater
biomass of down logs on higher surfaces was a con-
sequence of higher stand densities. Forests of
Surface 6 were not readily distinguished from those
of Surface 5.
31
Tree ages only roughly bracket the age of georaor-
phic surfaces. Oldest trees on Surfaces 2 and 3
may well date the time the geomorphic surface was
formed as a fresh substrate for vegetation estab-
lishment. Old-growth trees on higher surfaces
simply provided minimum estimates of surface ages.
Estimates of minimum ages of Surfaces 5 and 6 will
be improved with better understanding of stand
development and age and decay rate of pieces of
dead wood.
The mapped reach and adjacent areas examined only
in reconnaissance revealed some consistent patterns
of backwater channels and zones of addition and
accumulation of large organic debris. Whole trees
fell into the river where Surface 4 and higher sur-
faces were undercut on the outside of bends in the
river. These trees accumulated at the heads of
the downstream gravel bars (Surface 1) which are
persistent sites for accumulation of large debris.
In many instances, these debris accumulations
regulated water movement into high water channels
that occurred regularly along the back edges of
Surfaces 1, 2, and, in some cases, 3 on the inside
of bends in the river.
Many small tributary channels flow directly toward
the river over Surfaces 5 and 6, but then turn
downstream and parallel the river by flowing along
the back edge of Surfaces 3 or 4. The tributary
stream in the east portion of figure 1 took such
an indirect route. The net effect was to increase
the area of low gradient, valley-floor tributary
streams. The cause of this channel pattern was
not clear. The back edges of many surfaces were
wet areas, and surveys across the valley floor
revealed some tendency for surfaces to slope away
from the main channel (fig. 2). This could be a
product of a type of levee formation due to pref-
erential accumulation of sediment on the margin of
the surface along the river during periods of over-
bank flow. Low areas at the back edges of Surfaces
3 and 4 also might originate by the same processes
that form and maintain high flow channels along
the back edges of Surfaces 1 and 2. These types
of gravel bars with high centers in the axis of
the main channel have been observed in other
sediment-laden rivers (R. J. Janda, pers. comm. ) .
In some instances along the South Fork of the Hoh,
accumulations of large organic debris on the prows
of gravel bars may aid sediment accumulation along
the axis of the bar and direct high flows into a
channel along the back edge of the bar. As the
river continually changes its course, these high
water channels may be largely abandoned by the
main stem only to be occupied later by a tributary
stream.
INTERACTIONS OF FLUVIAL PROCESSES, LANDFORMS , AND
VEGETATION
Interactions between high flow events and forest
vegetation vary from one geomorphic surface to
another (fig. 3). Flooding by the main stem
affected Surfaces 1 through 4. This inundation
might affect plant community composition. One
effect was localized deposition of fine sediment
which might provide seed bed for species such as
alder that otherwise might not become established
in the stand. Surface 3 and higher surfaces were
subject to bankcutting while surface scour was the
more important erosion process on lower surfaces.
Floated large organic debris has both positive and
negative effects on live vegetation. Debris
carried by flood flow can severely batter living
plants on lower surfaces. Stabilized, large, down
debris provided protected sites where alder and
other pioneering species became established. Once
established, living vegetation in turn began to
stabilize geomorphic surfaces by developing root
systems and reducing water velocity by the flow
resistence of stems.
Some of these interactions between fluvial proc-
esses and vegetation could be interpreted from
analysis of the alder thicket ("camp thicket") on
Surface 2 along transect A-A' in figure 1. This
thicket appeared to have developed after floods in
1962, 1966, and 1968 (fig. 4). Tree ages varied
over more than a decade, suggesting occurrence of
repeated disturbance and opportunity for establish-
ment. The main body of the stand was protected by
several large logs partially buried in sediment
(fig. 5). Alder stems in bordering areas not pro-
tected by the down logs had been repeatedly and
heavily abraded by floating organic debris and
moving bedload sediment. The major down logs pro-
tecting the thicket and trees in the thicket itself
created a localized quiet water environment where
fine sediment was deposited during high flows.
This process, coupled with litter production by the
stand, accelerated soil development and growth of
the stand. The large, down debris helped the stand
reach a stage of structural development where it
could better withstand most floods.
Age analysis of sapling and small trees at the camp
thicket and other sites in the mapped area provided
additional insight into the role of flooding on
vegetation establishment (fig. 4). The camp-alder
thicket appeared to postdate the flood of November
1962, but the broad spread of ages indicated that
subsequent high flows provided new opportunities
for establishment. Some other areas mapped as Sur-
face 2 had trees nearly 30 years old which could
have been established after the second highest flow
on record in November 1949. Trees sampled on Sur-
faces 3 and 4 had a broad range of ages, some of
which clustered following years with high flows
(fig. 4). The group of 23- to 28-year-old trees
on Surface 4 appeared to postdate the high flow of
November 1949. The forest floor of this surface
was well covered with litter and herbaceous vegeta-
tion, suggesting that establishment of these alder
trees may have occurred on overbank deposits.
32
STABILIZATION
BY
GEOMORPHIC SURFACE
Figure 3. — Types of effects of fluvial processes
on geomorphic Surfaces 1 through 5.
In general, the lower geomorphic surfaces are re-
peatedly affected over a period of at least several
centuries by high flows which damage vegetation
and create new opportunities for establishment.
Forests on Surface 4, for example, contained trees
over 250 years old, and these stands were still
affected by flooding. Consequently, forests on
these surfaces were not simple single-aged stands
dating from single floods.
AQUATIC HABITAT
Geomorphic processes had created and maintained
four broadly defined classes of aquatic habitat in
the valley. The main river channel was character-
ized by fast, turbid water and shifting channel
position. Glaciers at the river's head are a
source of abundant silt and clay-sized sediment
during the spring, summer, and fall months. Ri-
parian vegetation had only moderate influence over
the wide river, and overall abundance of large or-
ganic debris was low relative to the other three
classes of stream environment.
The broad valley floor allowed development of a
variety of back-water and high-flow channels we
term "off-channel" sites. These sites ranged from
ephemeral high flow channels to secondary river
channels that carried water much of the year. In
some cases, log jams regulated flow into off -channel
areas. Compared with the main channel, flow veloc-
ity was moderate, and large organic debris and ri-
parian vegetation were more important. However,
this environment was quite varied with flow condi-
tions ranging from fast and turbid to slow and
clear to no flow at all.
Tributary streams originating from springs at the
back, edges of valley-floor surfaces or from streams
draining the valley walls were termed "terrace
tributaries". These streams were typified by low
gradients, quiet, clear water, and strong influence
of surrounding forest vegetation.
CAMP THICKET
• Alnus rubra
• Piceo sitchensis
Q
111
.kill.
oxxxxxxxx
XXXXXXIfX <
OTHER SURFACE 2 SITES
<
X
X
bJ
Id
or
S58 xP$m5S xxxxSxft**
i-
li.
o
SURFACE 3
cc
Id
CD
2
o 2 X
X X ex X OX X
XX o oX XX XXX XXoXXXoo X X XX oXoXX o
. 5 5
2
SURFACE 4
o XXoXX
o ooo e XXXXX o o
o o XXX o XOXXXSlX XoXooX o I o o
10 20 30 40 50 60
70
tttt tt
TREE AGE (yr)
(5) (3) (6)
ft t t t t t
t
(0(2) (4)
(3X8) (6) (2) (7) (5) (4)
(1)
12/4/75
1/16/74
12/26/72
3/5/72
1/19/68
12/13/66
1 1/19/62
1/13/61
1 1/3/55
11/26/49
2/7/45
11/5/34
12/12/21
0*
o
1965- 1977
Peok Flows
(rank)
Hon R.
1909-1964 Peak Flows (rank)
Hon R.,near Spruce
at Hwy. 101
Figure 4. — Age distribution of alder and spruce
saplings and small trees on Surfaces 1 through 4
and peak flows since 1909. Peak-flow data for
1927-57 for Hoh River and earlier peaks for
Quinault River from Bodhaine and Thomas (1964).
More recent data from annual U.S. Geological
Survey reports of Surface Water Supply of the
United States, Part 12, and Water Resources Data
for Washington.
Figure 5. — "Camp" alder thicket on Surface 2
located on transect A-A' (see figs. 1 and 2).
SAMPLED TREES loge
* P/ceo silchensis
* Alnus rubra
1 I Alnus thicket
r I Open orea, scattere
GRAVEL BAR
Mopped by S Lewis ano K Luehessa 9/76
33
VALLEY WALL
TRIBUTARIES
TERRACE -
TRIBUTARIES
V RIVER
OFF CHANNEL
MAIN RIVER CHANNEL
Figure 6. — Four types of stream environments.
Valley-wall tributaries were rigorous environments
with steep gradients and flashy flows. Since these
streams are not glacier fed and heavy forest vege-
tation on the valley walls minimizes soil erosion,
stream water was clear except during periods of
high flow. Forest vegetation strongly influences
valley-wall tributaries, particularly in the case
of large organic debris which forms a stair-step
profile along streams. Falls and plunge pools
formed in this fashion dissipate stream energy,
store sediment, slow removal of fine organic detri-
tus which is the food base of the aquatic ecosys-
tems, and shape microhabitats within the stream.
The influences of forest vegetation on aquatic
ecosystems and geomorphic forms and processes
increased across this range of stream types from
main river channel to valley-wall tributary.
Structure and productivity of aquatic communities
varied across these types in response to geomor-
phic factors and degree of forest influence (see
Sedell et al. and Ward et al. in this report).
CONCLUSIONS
Geomorphic features set the stage for development
of terrestrial and aquatic ecosystems of the South
Fork Hoh River. Geomorphic processes regulate the
types and rates of ecosystem development. Struc-
ture, age distribution, abundance of dead wood, and
other characteristics of forest communities varied
from one geomorphic surface to another in response
both to the time period available for forest de-
velopment and to degrees and types of flood influ-
ences. Successively higher surfaces were less
influenced by fluvial processes of the main river.
For example, vegetation on Surface 2 was subjected
to a variety of destructive and beneficial influ-
ences of high flows, while higher surfaces were
mainly affected by bankcutting and overbank depo-
sition of sediment. Alder establishment in spruce
forests on Surface 4 may have occurred on seedbed
provided by deposits of fine overbank sediment.
Thus, this surface appeared to have been affected
by fluvial processes for at least several centuries.
Geomorphic processes have created and maintained
four distinctive types of aquatic environments in
the valley (fig. 6): (1) the main river channel
with fast, turbid water and only moderate influ-
ence of forest vegetation; (2) off-channel areas
of the main stem with flows that vary from fast to
quiet and turbid to clear and may dry up part of
the year; (3) terrace tributaries characterized by
quiet, clear water; and (4) valley-wall tributaries
that have fast, generally clear water. Forest
influences increase across this array of stream
types.
Understanding geomorphic setting, both in terms of
landforas and processes, is essential to interpret-
ing terrestrial and aquatic ecosystems of the Hoh
valley.
LITERATURE CITED
Bodhaine, G. L. , and D. M. Thomas.
1964. Magnitude and frequency of floods in the
United States. Part 12. Pacific slope basins
in Washington and upper Columbia River basin.
U.S. Geol. Surv. Water-Supply Pap. 1687. 337 p.
Crandell, D. R.
1965. The glacial history of western Washington
and Oregon. In The Quaternary of the United
States. p. 34T-353. H. E. Wright, Jr., and
D. G. Frey, eds. Princeton Univ. Press, Prince-
ton, N.J.
Heusser, C. J.
1974. Quaternary vegetation, climate and glacia-
tlon of the Hoh River valley, Washington. Geol.
Soc. Am. Bull. 85:1547-1560.
34
Ecology and Habitat Requirements of Fish Populations in South Fork Hoh River,
Olympic National Park
J.R. Sedell, PA. Bisson, J.A. June, and R.W. Speaker
ABSTRACT
Four distinct running water habitats are defined and examined on the South Fork Hoh River — main river
channel, river off-channel areas, terrace tributaries, and valley wall tributaries. Species compositions,
densities, and total fish biomasses are distinctly different for each habitat examined. Habitat formed by
the main river channel and its tributaries is controlled by the valley terrace structure and the modifying
effects of large woody debris. Without large wood, spawning and rearing habitat quality would be poorer,
even in the large channel. Virtually all rearing of salmonid fish occurs in river off-channel areas and
tributaries. The main channel is used mainly for spawning and migration. Fish densities and biomasses
are highest in streams along the valley floor. Alteration of these areas will have greatest impact on
fish production.
J. R. Sedell, U.S. Department of Agriculture,
Forest Service, Forestry Sciences Laboratory,
3200 Jefferson Way, Corvallis, Oregon 97331; P. A.
Bisson, Weyerhaeuser Company, 505 North Pearl
Street, Centralia, Washington 98531; J. A. June,
Department of Fisheries, University of Washington,
Seattle, Washington 98195; and R. W. Speaker,
Department of Fisheries and Wildlife, Oregon State
University, Corvallis, Oregon 97331.
35
CLEMSCN UNIVERSITY LIBRARV
SOUTH FORK HOH RIVER
FISH SAMPLE SITES
WAIN RivER
OFF CHANNEL
6 -
TERRACE
11-
LOWER VAILE* V
'All 15-
UPPER vALLE* V\
AU IT-
Figure 1. — Location of fish sample sites on the
South Fork Hoh River and the tributaries.
INTRODUCTION
Stream ecosystems are adaptations to interacting
geologic forces (hydrology and fluvial forms and
processes) and biotic modifiers (riparian vegeta-
tion and large downed trees). Diversities of flow
velocities, organic input quality and quantity,
storage and processes of organic materials, and
habitat types are all directly affected. The
major interest in the South Fork Hoh River stems
from the fact that the headwaters flow from a
glacier, resulting in a channel that has many
migrating point bars and a continual supply of
glacial flow sediments in suspension. At present,
stream ecologists have no benchmark descriptive
data for naturally sediment-rich streams. We know
very little about aquatic habitats and how they
are formed in pristine sediment-rich river valley
systems. For these reasons, a description of this
river in the Olympic National Park provides a
needed and significant baseline for stream
ecologists and fisheries managers in the Pacific
Northwest.
Our part of the South Fork Hoh River "pulse study"
was to examine fish populations in relation to the
four previously defined habitats (main river
channel, river off-channel areas, terrace tribu-
taries, and valley wall tributaries (see Swanson
and Lienkaemper, and Ward and Cummins in this
report)). Species compositions, densities, and
total fish biomasses were defined for each habitat
type. With each habitat type, we obtained a
description of the pools, channel stability, and
role of large woody debris in the formation and
stability of the habitat.
MATERIALS AND METHODS
A beach seine was employed for sampling salmonid
populations in the main stem South Fork Hoh River.
A 600-volt backpack electric shocker was used to
sample off-channel and tributary sites. A single
pass method was used on 5 main stem sites and a
two pass removal method on 13 off-channel and
tributary areas (fig. 1). Fish collected from
each site were anesthetized with MS 222 (tricaine
methanesulf onate) , identified to species, measured
for fork length, and allowed to recover before
release. Fish biomass estimates were calculated
from length-weight relationships determined pre-
viously from other watersheds. Density, biomass,
and species distribution were related to each habi-
tat parameter and important relationships noted.
Each site was measured for length and width, and
wetted surface areas were computed. Debris
obstructions were counted; and their role in bank
stability, fish cover, and flow deflection was
noted. The stability of each site was determined
using USDA Forest Service "Stream Reach Inventory
and Channel Stability Evaluation," (Pfankuch 1975)
and categorized as very stable, stable, or
unstable. The quality of pools in each site was
determined using a modified Duff and Cooper (1976)
technique assessing pool volume and cover.
A survey of spring-summer chinook salmon spawning
sites was conducted during September and October.
Foot surveys were the primary method; helicopter
surveys were utilized during peak spawning
activity.
Gravel samples were taken at two sites on the main
river and one tributary site. Likely spawning
gravels were sampled with a McNeil cylinder and
Koski Plunger. A total of 25 samples was analyzed
volumetrically for percent material less than
.85 mm.
HABITAT DESCRIPTIONS
Main River Channel
The main river was wide and shallow (table 1).
Wet widths ranged from 8 to 10 meters in summer to
20 to 40 meters in winter. The main channel
meandered within a wide channel of exposed gravel
bars that averaged 100 m wide. The large cobble
substrate was very unstable, gravel bars being
formed and destroyed continuously. Some bank
cutting was evident along the steep south valley
side slope. The water was turbid due to suspended
glacial material. Organic material transfer and
storage was low. Edges of the main channel
accumulated sediments, but riffle areas were
relatively clean. The channel gradient was 2 to 3
percent and mainly riffles and deep runs with some
pools associated with debris. Riparian vegetation
did not significantly influence the course of the
river; however, bank cutting caused inputs of
large woody debris which would accumulate on bars
and cutting edges to deflect the river flow (see
Swanson and Lienkaemper in this report). Debris
accumulations provided little fish cover in the
main river, but diverted water through off-channel
overflow areas.
36
Table 1 — Physical characteristics of the major aquatic habitats in the South Fork of the Hoh River,
autumn 1978
Habitat type
Physical characteristics
Stability
Debris
Pool
Riffle
Site number
Main river sites
(1,2,3,4,5)
Poor stability, winter and spring
floods cause cutting and deposition.
Debris collected in small
jams at cutting areas on
bends in river, stabilizes
banks and deflects flow.
Percent Percent
20
80
Off-channel sites
(6,7,8,9,10)
Terrace tribs.
sites (11,12,
13,14)
Lower valley
wall tribs.
sites (15,16)
Upper valley
wall tribs.
sites (17,18))
Good stability except during
extremely high main river flows.
Very good stability low gradient,
debris-protected banks.
Very good stability high gradient,
boulder- and debris-stabilized
banks.
Excellent stability high gradient
steep banks, boulder-formed plunge
pool and falls.
Debris accumulations on main 70 30
channel creates and maintains
most off-channel area. Indi-
vidual pieces offer fish cover.
Individual pieces reduce 80 20
cutting of banks and offer
fish cover.
Individual pieces help 40 60
stabilize banks and form
some plunge pools.
Large individual pieces 90 10
offer some bank stability
and with boulders form plunge
pools and fish cover.
Off-Channel Areas
Channels subsidary to the main river were located
within the active exposed lower flood plain. Some
were caused by debris accumulations on bars in the
main channel, river flow was diverted, and a gravel
berm was created downstream from the debris. Water
percolated through the gravel berm and debris to
create a side channel between the gravel berm and
the bank opposite the main river flow channel.
Other off-channel areas were intermittent overflow
channels that received ground water from the main
river and nearby terrace. Most were subject to
direct flows during freshet periods; others became
completely isolated during summer low flow periods.
Flow velocities are lower than the main river, and
water percolated through berm gravels carries
reduced suspended sediment. Organic input from
terrace vegetation and overflow accumulations from
main river floods collected in off-channel pool
bottoms. In the absence of heavy shading and the
scouring effects of suspended glacial material,
algal growth was promoted. High insect production
occurred in these organically rich areas (see Ward
and Cummins in this report). Woody debris and
undercut bank vegetation provided cover for fish
and created pool areas. Downstream tailouts
offered good spawning locations.
Terrace Tributaries
Terrace tributaries result from spring networks on
the flat valley flood plain and from tributaries
draining the valley side slopes and continuing
across the terraces to the main river. Many
terrace tributaries paralleled the secondary river
channels that cut through the lower terrace areas
within the flood plain before emptying into the
main river. These streams were very stable and
had low gradients, slow velocities, and channel
widths from 1 to 5 m. They were composed predom-
inately of pools and short sections of riffle.
Pools accumulated large amounts of riparian leaf
litter from the dense forest canopy, thus produc-
ing abundant aquatic insects (see Ward and Cummins
in this report). Pool substrate was primarily
fine sediments, although riffles were relatively
clean. Banks were stablilized by live vegetation
and downed woody debris. Debris and undercut
banks provided excellent fish cover.
Lower Valley Side Wall Tributaries
Side slope streams originated from runoff on the
steep valley walls. The lower ends of these
streams flowed on the upper terrace areas from the
north valley walls and directly into the main Hoh
River from the south valley walls. These streams
were typified by high gradients with alternating
sections of riffles and plunge pools over woody
debris and boulder obstructions. The substrate
ranged from fine materials deposited above some of
the debris to large boulders and bedrock in the
37
plunge pools. Lower valley side wall tributaries
had clear water and high flow velocities. Organic
material from the riparian vegetation was trans-
ported downstream or retained by the debris and
boulders. The banks were steep but stable due to
deeply embedded large boulders and debris.
Productivity of algae and aquatic invertebrates
was lower than other streams because the dense
forest canopy limited light entry and scour forces
of the rapidly moving water created considerable
shear stress.
Upper Valley Side Wall Tributaries
Upper side slope streams had very steep gradients
and high velocities. They flowed over a series of
stair-step pools and cascades set up by boulders
and large downed trees. The substrate was large
cobbles, boulders, and bedrock, with some gravel
accumulation at the tails of pools and behind
embedded debris. The water ran clear and the high
velocity transported downstream any organic
material not retained by the boulders and debris.
The steep banks with high boulder content and
large downed trees maintained the stable
riffle-cascade-pool nature of the channel. Low
benthic algal production from reduced light
penetration resulted in low densities of aquatic
insects, except for some wood gougers.
THE ROLE OF LARGE ORGANIC DEBRIS IN THE
MAINTENANCE OF FISH HABITATS
Large organic woody debris in streams of the
spruce-hemlock forest have profound effects on
channel form and fluvial processes, particularly
in small sized streams. Woody debris (10-cm
diameter) plays different roles in each of the
habitat types described. We examined the extent
to which debris intervened in the stream channel.
We grouped debris interventions in the channel
into three groups, depending on the extent of
direct influence within the channel width. The
groups were influences of one-third to two-thirds
of the channel width and complete channel dams,
bridges, or other direct interventions.
On large, main channels, the woody debris
accumulated at the head of gravel bars and often
regulated water movement into off-channel areas
(see Swanson and Lienkaemper in this report).
Even the largest trees seldom crossed the entire
channel parallel with the flow or in accumulations
below the curving bend of the river. We found
large debris intervened up to one-third the
channel width 10 times, two-thirds the channel
width once, and crossed the main channel twice
along 900 m of main river stem. Half of the
debris interventions influencing up to one-third
the channel were at the head of off-channel
areas. The root wads of single trees in the main
channel that were parallel to the flow sometimes
deflected the flow toward other debris or
boulders. These convergences of flow created
pools, and the tallouts of these pools provided
excellent spawning habitat for chinook salmon.
The importance of a large downed tree to fish
habitat in the stream channel cannot be minimized,
even though it may intervene less than one-third
of channel width. Their role as flow deflectors
contributed to the diversity of flow velocities
that helped maintain spawning areas free of the
fine fluvial sediments which could smother
incubating fish eggs.
Off-channel habitat responding to intervention by
large woody debris were highly variable. Often
the flood flows had resulted in accumulated debris
along the edge of the main channel, deflecting the
flow and producing large backwater pools or
alcoves. These off -channel alcoves often were
important juvenile coho and steelhead rearing
areas. These kinds of habitat were created by
100 percent intervention of the debris. The
secondary and tertiary channels frequently took
the form of small off-channel streams. The woody
debris intervening in these channels was derived
primarily from sources upstream and was deposited
through a major storm event. For every 100 m of
off-channel area, there were 12 interventions of
wood one-third of the channel water or less, two
interventions influencing up to two-thirds of the
channel, and four interventions which dammed or
influenced the complete channel width. About
60 percent of these channels were high quality
pools formed by debris and were primary rearing
areas for large juvenile coho and steelhead.
Terrace tributaries had accumulated large wood
from their surrounding forest. These were predomi-
nately pool environments except at their lower
reaches where they merged with the main channel.
For every 100 m of terrace tributaries channel,
large woody debris less than one-third channel
width category intervened nine times, four times
for influences from one-third to two-thirds
channel width, and three times the entire channel
width was dammed or influenced directly. These
pieces of wood were quite stable, and supported
nurse trees and dense moss communities.
The valley wall tributaries had a stepped profile
created by sediments stored behind large wood in
the channel. Pools usually were created by debris
and provided primary rearing areas for many of the
river system's cutthroat trout. In general, for
every 100 m of this stream type, we found 11 wood
interventions influencing the channel width one
third or less; 5 times the entire channel was
dammed or directly influenced.
In general, the main channel and off channel areas
utilized trees and large pieces of wood that
originated upstream from where the accumulations
were found. The forest along terrace tributaries
and valley wall tributaries contributed the wood
usually found in these streams. Debris was a
major contributor to fish habitat for both
spawning and rearing requirements of the different
fish's life cycles. Although we tend to ignore
debris influence on the physical channel of large
rivers, its role in forming and maintaining
anadromous fish habitats is very important
regardless of size of streams. Without large
trees being transported by the main channel, the
very productive off-channel areas would maintain
levels of invertebrate and fish densities and
biomasses much lower than they now do.
38
Large trees or wood in streams do not have to dam
a stream channel completely to have a major
influence on fish habitat. The majority of debris
intervening on channels influenced only one third
or less of the channel width. This was enough to
create diverse stream velocities, pocket pools,
and cover, which resulted in stable and diverse
fish habitat conditions.
LIFE HISTORIES OF SALMONID FISHES IN THE
SOUTH FORK HOH RIVER
Mountain Whitefish - Prosopium williamsoni
Whitefish were taken only from the main river
channel itself; no specimens were collected from
off-channel or tributary areas. Within the main
channel, whitefish probably frequented deeper runs
and pools where they fed upon aquatic inverte-
brates. Car, Clemens, and Lindsey (1967) report
that this species spawn in October and November,
their eggs hatching around March. The mountain
whitefish is not anadromous and apparently
completes its entire life cycle within the main
river channel.
Dolly Varden - Salvelinus malma
Most individuals of this species were collected
from the main river channel, but a few juveniles
were taken from off-channel areas. Larger adult
fish have been known to feed extensively on small
fish, including migrating smolts. Immature Dolly
Varden feed mainly on aquatic invertebrates. This
species is known to be anadromous in some cases,
but we did not determine if the population in the
South Fork Hoh was anadromous or resident. Dolly
Varden are occasionally caught by anglers fishing
the main river channel. With the exception of
some off-channel rearing, this species also is
confined to the main stem.
Coastal Cutthroat Trout - Salmo c la rki
Cutthroat were captured in the tributaries, the
majority of individuals being taken from the upper
valley wall tributaries where they were collected
from plunge pools and undercut banks and logs.
All life history stages feed on aquatic and
terrestrial invertebrates, although larger adult
cutthroat will feed on small sculpins. Sea-run
cutthroat populations occur in most coastal river
systems of Washington, but scale analysis of
samples collected from South Fork Hoh River
specimens during this study indicated that the
population was composed entirely of resident fish.
This species spawns in the spring, probably in the
valley wall tributaries. Emergence of young-of-
the-year occurs in June and July. It is safe to
conclude that cutthroat trout make up the great
majority of the sport fishery in the valley wall
tributary systems, and it is likely that more
cutthroat are caught in the South Fork Hoh River
drainage than any other sport fish.
Steelhead Trout - Salmo gairdneri
Juvenile steelhead were taken from the main
channel, off-channel, terrace tributaries, and
lower valley wall tributaries. No adult steelhead
were captured, although they were known to be
present in the main stem. Steelhead are
anadromous rainbow trout; and two distinct runs
are known to occur in the South Fork Hoh River —
one run takes place in summer and a second run
occurs in winter. Although the timing of the two
runs is separate and individuals belonging to the
runs are known to have certain genetic differ-
ences, spawning times show considerable overlap.
Depending upon time of entry into fresh water,
spawning can occur from December to May, although
peak spawning activity probably takes place in
February and March. Steelhead fry are usually out
of the gravel by June, and most rearing occurs in
off-channel and valley wall tributary areas. Size
distribution of juveniles suggested that most
steelhead spend 2 years in the South Fork habitat
before smolting and migrating to sea. A notable
sport fishery for adult steelhead exists on the
main stem, and some legal-sized juveniles are
probably caught in the tributaries along with
cutthroat.
Coho Salmon - Oncorhynchus kisutch
Juvenile coho were found mainly in river
off-channel and terrace tributary areas, where
deep slow-moving water occurred. At the time of
the study (late September), adult coho migration
into the South Fork had not yet taken place; the
bulk of the run was expected in November. Adult
spawning probably occurs in both the main stem and
valley wall tributaries; but juveniles were
conspicuously absent from these areas, having
moved to pools adjacent to the river and into the
terrace tributaries for rearing. Young coho were
heavily dependent upon terrestrial invertebrates
for food, particularly during summer months. They
usually spend over a year in fresh water before
smolting in spring and returning as adults after 2
years at sea. The contribution of coho to the
South Fork sport fishery probably is small.
Chinook Salmon - Oncorhynchus tshawytscha
No juvenile chinook were taken at any sample
loction in the South Fork Hoh River; apparently
this species does not rear there in early autumn.
Like steelhead, there are two adult runs in the
river — one in late spring and the other in autumn.
We observed adults spawning in the main river
channel and in the lower reaches of the larger
tributaries. These were presumably spring-summer
run fish. Approximately one-third of the total
spring-summer chinook run entering the Hoh River
drainage spawns in the South Fork. Of these,
about two-thirds spawn in the main and secondary
river channels and one-third spawn in the terrace
tributaries and the lower valley wall tributaries
(table 2). Juvenile emergence occurs in late
winter; and young-of-the-year move downstream to
rear in other parts of the system, principally in
the estuary.
39
Table 2 — Spring-summer chinook salmon spawning
Percent River channel
of total and river Terrace Valley wall
redd count off-channel tributaries tributaries
170
65
22
13
FISH POPULATION DENSITY, BIOMASS, AND GROWTH
The largest density and biomass of salmonids
occurred in the off-channel habitat (table 3).
Steelhead young-of-the-year (fry) represented
79 percent of the total density (fig. 2A and 2B) .
Coho salmon fry made up 19 percent of the total
density, but were larger than the steelhead fry
(table 4) and accounted for 56 percent of the
total biomass. Steelhead, although more abundant
than coho, comprised only 39 percent of the
biomass. Cutthroat trout and Dolly Varden
contributed little to salmonid density and
biomass. Sculpin density was one-fourth of the
total salmonid density, but sculpin biomass was
equal to that of the salmonids.
The terrace tributaries possessed the second
highest density of salmonids, yet density was less
than half of the off-channel areas. Coho were the
most abundant, and made up 76 percent of the total
density, followed by cutthroat trout, with 17
percent of the total density. Coho fry were
smaller in the terrace tributaries (table 4) than
in the off-channel areas but, because of their
higher density, accounted for a biomass nearly
equal to that of coho in off-channel areas. The
coho biomass was 62 percent of the total salmonid
biomass in the terrace tributaries. Cutthroat
trout averaged 35 percent of the total salmonid
biomass. Steelhead fry were both small and rare
in the terrace tributaries. Sculpins in terrace
tributaries were smaller than those of off-channel
areas but were twice as abundant.
Upper valley side wall tributaries had a total
salmonid density lower than both off-channel area
and terrace tributaries; total salmonid biomass
was lower than the off-channel areas but higher
than the terrace tributaries. Cutthroat trout
represented 97 percent of the density and 92
percent of the biomass of salmonids collected.
Cutthroat biomass was relatively high, and the
wide range of sizes represented in the population
indicated that several year-classes were present.
A single steelhead yearling was the only other
salmonid captured in upper valley wall tributaries.
Sculpins were present, but their density and
biomass was less than a third of those found in
the off-channel and terrace tributaries.
Lower valley side wall tributaries had lower total
salmonid density and biomass than the off-channel,
terraces and upper side wall tributaries. Steel-
head trout fry dominated both density and biomass.
The mean length of steelhead fry exceeded that
found in the terrace and off-channel areas. Sev-
eral coho salmon and cutthroat also were captured.
Cutthroat trout made up only 13 percent of the
total salmonid density but comprised 31 percent of
the biomass due to two large yearlings in the
sample. Sculpin density and biomass were lower
than all the other habitat areas.
The main stem South Fork Hoh River had the lowest
salmonid density and biomass of all habitats. A
total of seven steelhead trout (three fry and
three yearlings), one Dolly Varden, and one moun-
tain whitefish were captured. No sculpins were
collected. Total salmonid density was estimated
to be less than .001 fish/m^, and biomass was
less than .01 g/m'
Mature adult chinook salmon
were observed migrating in the main stem to spawn-
ing sites in the off-channel areas and lower ter-
race tributaries, but none was collected in the
seine samples. These low values probably reflect
inefficient sampling techniques. While we believe
the main river channel to possess fewer fish, the
values in tables 3 and 4 and figure 2A should be
considered tentative. Accurate main stem esti-
mates will require a more effective large river
sampling program.
Generally, the densities of fish in the South Fork
Hoh River were lower than reports for other
streams on the Olympic Peninsula outside of the
Park. Coho density in off-channel and terrace
tributaries were similar to values reported for
some tributaries to the nearby Clearwater River
(Edie 1975). Cutthroat trout density on the upper
and lower valley wall tributaries were less than a
third of that reported for lower gradient headwater
streams in the Clearwater and Bogachiel Basins
(Lestelle 1978, Martin et al. 1978).
Gravel samples taken in the main channel and lower
valley wall tributaries (J. Cederholm, pers.
comm. ) were similar in percentage of fine sediment
to pristine watershed in the Clearwater River.
This is important, since the Clearwater is not a
glacier-fed stream.
The importance of off-channel ponds in the success-
ful rearing of coho salmon smolts in a river system
has been documented by Peterson (1979). While our
brief study did not include any ponds of the size
and depth studied by Peterson, we feel that the
numerous flood-influenced off-channel pools and
alcoves along the river's border provided the bulk
of the juvenile anadromous fish rearing areas.
MANAGEMENT IMPLICATIONS
Biological productivity of the South Fork Hoh
River is largely dependent on stable terrace
stream networks and valley wall tributaries.
These productive zones can be protected from abuse
by proper campground and trail placement. Major
trails should avoid paralleling terrace tribu-
taries; they would be better placed on lower
valley walls. Avoiding terrace tributary areas
for locating campgrounds also will help protect
the streams. Existing road networks could be
re-examined to determine if biologically diverse
and productive areas have been cut off from the
floor influence of the main channel and If fish
passageways to and from off-channel rearing sites
have been blocked.
40
Table 3 — Density and biomass of fish species collected in stream habitats of the South Fork Hoh River,
autumn 1978
Coho
Steelhead Cutthroat Dolly Varden Total salmonid Sculpln Total fish
Habitat type #/m2 g/m2 #/ra2 g/m2 #/m2 g/m2 #/ra2 g/m2 #/ra2 g/m2 #/m2 g/ra2 #/m2 g/m2
Main river
—
—
.003
.01
Off-channel
.070
.33
.286
.23
Terrace
tributaries
.118
.26
.010
.01
Lower valley
wall trib-
utaries
.003
.01
.044
.08
Upper valley
wall trib-
utaries
—
—
.002
.04
.001
.026
.007
.065
.02
.15
.04
.48
.001
.003
.01
.02
.004
.364
.154
.053
.067
.01
.59
.087
.42 .156
.13 .029
.52 .042
.59
.47
.09
.13
.451 1.18
.310 .89
.082 .22
.109 .64
*Sculpins were not sampled in the main river; therefore, we omitted computation of total fish density and biomass.
Relative Biomass (g/m2)
E3 Coho salmon
□ Steelhead trout
□ Dolly varden trout
□ Cutthroat trout
Main
river O
Terrace tributaries
Off channel
Figure 2A. — Average biomass of salmonid fishes in
the South Fork Hoh River drainage system.
Figure 2B. — Average density of salmonid fishes in
the South Fork Hoh River drainage system.
Relative Density (#/m2)
U Coho salmon
El Steelhead trout
□ Dolly varden trout
0 Cutthroat trout
Main river •
Valley wall upper
tributaries
Lower r5j
Terrace tributaries
Off channel
Table 4 — Mean and range of length in millimeters of
four species of fish collected in five aquatic
habitats of the South Fork Hoh River, autumn 1978
Species
Habitat
type
Coho Steelhead Cutthroat Sculpin
Main river
125.0
(50-131)
Off-channel 73.6 42.0
areas (55-101) (31-70)
113*
Terrace 57.3 36.3 77.0
tributaries(40-85) (31-41) (51-138)
Lower valley 68.5 51.3 74.0
wall trib- (61-76) (33-67) (42-120)
utaries
Upper valley
wall trib-
taries
131*
79.0
68.7
(24-117)
57.0
(44-97)
55.0
(40-110)
53.0
(38-170) (41-95)
*0nly one individual captured.
41
The Olympic National Park provides a full range of
pristine river systems which can be used as a
bench mark for thousands of square kilometers of
adjacent altered river systems. There is a basic
need in the State of Washington and throughout the
Nation to improve both the systems we manage and
the system we manage with. Aquatic biologists
lack comprehensive knowledge of pristine systems
with relatively unmanaged fish population. The
information we collected is of the kind needed for
understanding the functioning of aquatic ecosys-
tems. Moreover, it allows us to begin comparing
the condition of pristine systems with the condi-
tion of similar, but altered, basins to develop an
understanding of the control mechanisms and stabil-
ity features of aquatic systems. An understanding
of pristine watersheds should lead to more meaning-
ful interpretation of the processes and effects of
both natural and artificial habitat alteration and
will also help promote more effective habitat
improvement programs.
ACKNOWLEDGMENTS
LITERATURE CITED
Carl, G. C. , W. A. Clemens, and C. C. Lindsey.
1967. The freshwater fishes of British
Columbia. Br. Columbia Prov. Mus. Handb. No. 5.
192 p.
Duff, D. A., and J. L. Cooper.
1976. Techniques for conducting stream habitat
survey on Natural Resource Land. U.S. Dept.
Interior, BLM Tech. Note 283. 72 p.
Edie, B. G.
1975. A census of juvenile salmonids of the
Clearwater River Basin, Jefferson County,
Washington, in relation to logs. M.S. thesis,
Univ. of Wash., Seattle. 86 p.
Lestelle, C. L.
1978. The effects of forest debris removal on a
population of resident cutthroat trout in a
small headwater stream. M.S. thesis, Univ. of
Wash. , Seattle. 85 p.
We are grateful to Phil Peterson, Steve White,
Jeff Cedarholm, Bill Wood, Vickie Era, Brian
Joannason, and Ray Palmason for technical
assistance. Our paper also benefitted from
discussions with Fred Swanson, U.S.D.A. For-
estry Sciences Laboratory, Corvallis, Oregon,
and Milt Ward, Ken Cummins, and Stan Gregory
of Oregon State University.
Martin, D. J., S. T. White, W. T. Foris,
J. A. June, and E. 0. Salo.
1978. Bear Creek streamside timber removal
study; Progress Report 1978. Progress Rept. ,
Fisheries Res. Inst., Univ. of Wash., Seattle.
61 p.
Peterson, P.
1979. The role of spring ponds in the winter
ecology and natural production of coho salmon on
the Olympic Peninsula, Wash. M.S. thesis, Univ.
of Wash. , Seattle.
Pfankuch, D. J.
1975. Stream reach inventory and channel
stability evaluation. U.S.D.A For. Serv.
Reg. 1, Missoula, Mont. 26 p.
42
Relationships Within the Valley Floor Ecosystems in Western Olympic National
Park: A Summary
Jerry F. Franklin, Frederick J. Swanson, and James R. Sedell
ABSTRACT
The major findings of the South Fork Hoh River research team emphasize interrelationships among components
of the Olympic valley-bottom ecosystems. Geomorphic structures provide the basic template for both terres-
trial and aquatic communities. Vegetation has significant reciprocal impacts on geomorphic processes, how-
ever, and is a major element in formation of the most productive aquatic habitats. The South Fork, pulse
illustrates accomplishments possible with intense, short-term interdisciplinary research efforts and the
valuable functions National Parks can perform as benchmark areas to compare with exploited land systems.
Other papers in this report consider geomorphology, forest communities, Roosevelt elk, and aquatic habitats
and communities. There is some tendency to lose sight of the entire valley-bottom ecosystem in these more
component-oriented presentations, however.
The objective in this paper is to racapitulate the major findings of the South Fork research pulse with an
emphasis on interrelationships among various components and processes. We also suggest broader implica-
tions for Park management and point out how the Park is serving as a control or baseline site for inter-
preting man's impacts on adjacent managed landscapes and establishing guidelines for improved management.
National Parks provide the rare opportunities to study natural, undisturbed, valley-bottom forests and
river ecosystems.
Jerry F. Franklin and Frederick J. Swanson, U.S.
Department of Agriculture, Forest Service, Pacific
Northwest Forest and Range Experiment Station,
Forestry Sciences Laboratory, 3200 Jefferson Way,
Corvallis, Oregon 97331; and James R. Sedell,
Weyerhaeuser Company, Forest Research Laboratory,
505 N. Pearl Street, Centralia, Washington 98531.
43
GEOMORPHIC RESULTS
Landforms can be viewed as the template on which
the terrestrial and aquatic communities of the
valley-bottom ecosystems develop. While geologi-
cal processes determine the initial conditions,
biological processes are significant modifiers.
Vegetation-geomorphic interactions are particu-
larly important in the cases of small streams and
river bars and terraces which are entirely the in-
terplay between vegetation and fluvial processes.
Woody debris creates some of the most conspicuous
vegetation influences on geomorphic processes.
Fluvial processes mobilize large amounts of woody
debris, particularly by undercutting and uprooting
trees on forested alluvial flats and higher ter-
races. Stabilized debris in the main river channel
is important in setting up gravel bars, protecting
pioneering vegetation from high flows and from buf-
feting by floated organic debris, and regulating
flow into river side channels, thereby creating the
especially productive off-channel aquatic habitat.
Woody debris in off-channel and tributary habitats
also provides physical stability (Swanson 1980), a
diversity of biological habitats, and an ecosystem
energy base both by retaining fine allochthonous
material and directly through decomposition (Cum-
mins 1980).
Four major categories of aquatic habitat are iden-
tifiable in river valleys: the main river channel
with fast, turbid, silty water; river off-channel,
such as side channels partially isolated from the
main stream; terrace tributary, which is low gra-
dient and generally carries clear, slow-moving
water; and valley wall tributary, typically a high
gradient stream with clear, fast water. Geomor-
phic, hydrologic, and vegetation factors combine
to determine the basic types and arrangements of
aquatic habitats.
AQUATIC RESULTS
Off-channel and terrace tributary habitats are of
overwhelming importance for productivity of aquat-
ic ecosystems of the South Fork and similar, broad,
alluviated valleys in western Washington. These
relatively protected sites have abundant, diverse
food resources for both invertebrates and fish.
Coho and cutthroat trout use these areas for rear-
ing. Terrace tributaries provide important shelter
when the main channel is in flood.
Glacial silt limits productivity of some parts of
the main channel ecosystem, but not others. Pri-
mary production in fast water areas of the channel
is severely limited by the scouring action of silt
and fine sand being transported in suspension much
of the year. Silt deposition in the few quiet water
sites of the main channel prevents full utilization
by invertebrates of these potentially very produc-
tive areas. Spawning is not greatly reduced by the
fine sediment because hydraulic conditions prevent
excessive accumulation in the major spawning sites
where pools tail out into the heads of riffles.
Productivity of valley wall tributaries is limited
by the extremes of high and low flows and the dense
conifer overstory which reduces primary production
by shading. Limited pool area and the difficulty
of moving from pool to pool along these high gradi-
ent channels constrain use of valley wall tribu-
taries by fish.
Large organic debris is an important factor in
shaping microhabitats in each of these types of
stream environments.
TERRESTRIAL RESULTS
The forest communities are strongly related to
landform or geomorphic surface. Alnus rubra
stands dominate youthful fluvial deposits and
mature Picea sitchensis-Tsuga heterophylla forests
occupy older, higher surfaces. In the study area,
different Picea-Tsuga communities are formed on
upper and lower terraces, although dominant trees
on both terraces are approximately the same age
and cannot be considered successionally related.
The vegetation-landf orm model proposed for the
main fork of the Hoh River (Fonda 1974) does not
fit the South Fork, which probably reflects his-
torical differences in timing and patterns of
forest disturbances such as floods.
Mature valley-bottom Picea sitchensis-Tsuga hetero-
phylla forests contrast with those found elsewhere
in the coastal P. sitchensis zone of the Pacific
Northwest (Franklin and Dyrness 1973). Stands,
especially those on lower terraces, are open with
relatively low density of above-ground biomass and
numerous openings of up to a hectare or more.
Picea is reproducing successfully, earning recog-
nition as a climax tree species on these sites.
While the relative success of Picea and Tsuga re-
production appear to have oscillated over the past
century, there is no evidence that either is going
to be replaced successionally. Grazing by Roose-
velt elk is a factor that may favor survival of
Picea reproduction over that of Tsuga, but grazing
is almost certainly not the sole cause of variation
in reproductive success. Wind and floods appear to
be the major environmental factors disrupting these
valley-bottom forests, while wildfire appears to
be inconsequential.
Coarse woody debris is an extremely important
structural feature of the terrace forests. Woody
debris occupies much of the forest floor and con-
tains large masses of carbon and nutrients. Re-
production of trees on older terrace surfaces is
confined almost exclusively to rotten logs and
associated stumps and root wads. Logs vary sig-
nificantly in their value as nurseries depending
upon log species, decay state, and terrace level.
Forest renewal is dependent on seedbeds of coarse
woody debris.
44
APPLICATIONS
CONCLUSIONS
The South Fork of the Hoh River appears to be the
archetype of the western Olympic Mountain river
valley ecosystem. Fluvial landforms and pro-
cesses, mature Picea-Tsuga terrace forests, and
valley-bottom aquatic habitats are well repre-
sented. The valley habitats have undergone mini-
mal modification from adjacent mountain sideslopes
and river tributaries in comparison with the other
four major river valleys of the western Olympic
Mountains. It therefore seems appropriate to
manage and utilize the South Fork valley as a
primary site for research on Olympic rainforests
and associated streams.
Recognition that off-channel and terrace tributary
habitats are aquatic hotspots and essential to
anadromous fish in the Olympic river valleys has
implications for managers inside and outside the
Park. Park managers should locate trails, roads,
and other developments so as to have minimum impact
on these features. Resource managers outside the
Park should appreciate the importance of providing
off-channel and terrace tributary habitats with at
least as much protection as the main channel. Re-
habilitation of such habitats may be essential in
areas where they have been destroyed by logging or
road construction activities.
The South Fork study demonstrates the use of a Na-
tional Park and Biosphere Reserve as a benchmark
site for scientific research and a control area for
adjacent manipulated landscapes. The only remain-
ing natural examples of river valleys in the coast-
al region are within Olympic National Park. The
knowledge gained in the pulse study has extended
our understanding of northwestern ecosystems to a
distinctive variant (the rain forest) and a larger
scale (river drainage).
Finally, the pulse in the South Fork of the Hoh
River demonstrates the numbers and types of data
that can be gathered by an interdisciplinary re-
search team in a short time span. A successful
project is based on substantial logistical plan-
ning and a balance between careful definition of
objectives and ample opportunity to pursue prom-
ising leads and for serendipitous discoveries.
LITERATURE CITED
Cummins, Kenneth.
1980. The multiple linkages of forests to
streams. In Richard Waring, ed. Proceedings,
20th Annual Biology Colloquium. Oregon State
Univ. Press. Corvallis, OR. In Press.
The significance of woody debris is further docu-
mented for both scientists and resource managers.
Debris can now be seen to play important roles in
larger streams and rivers, roles which must be
accounted for by land managers in programs of ri-
parian management and stream cleanup. The role of
woody debris as critical seed bed in coastal forest
types has implications on lands managed for timber
production as well as on lands reserved from devel-
opment.
Fonda, R. W.
1974. Forest succession in relation to river
terrace development in Olympic National Park,
Washington. Ecology 55( 5) :927-942, illus.
Franklin, Jerry F. , and C. T. Dyrness.
1973. Natural vegetation of Oregon and Washing-
ton. USDA For. Serv. Gen. Tech. Rep. PNW-8,
417 p. , illus. Pac. Northwest For. and Range
Exp. Stn. , Portland, OR.
Swanson, Frederick J.
1980. Erosion as an ecosystem process. In
Richard Waring, ed. Proceedings, 20th Annual
Biology Colloquium. Oregon State Univ. Press.
Corvallis, OR. In Press.
45
Introduction and Dispersal of Mountain Goats in Olympic National Park
Bruce B. Moorhead and Victoria Stevens
ABSTRACT
Although native to the Cascade Range of central Washington, the mountain goat, Oreamnos americanus, was
introduced to the Olympic Mountains of northwestern Washington by man prior to creation of Olympic National
Park. There is no historical evidence of goats in the Olympic Mountains before release of 11 or 12 from
Canada and Alaska between 1925 and 1929.
Dispersal of goats is traced in the last 50 years from release sites near the present north Park boundary.
Moving east and south the population apparently reached the southern edge of the mountains (50 miles
(80 km) distance) by about 1960. It is presently distributed throughout the Olympic Mountains in over
700 square miles (1 800 km ) . Ninety percent of the occupied habitat is in the Park.
Bruce B. Moorhead, Science and Technology Group,
Olympic National Park; and Victoria Stevens,
College of Forest Resources, University of
Washington.
46
INTRODUCTION
Through its management policies, the National Park
Service seeks to perpetuate and restore native eco-
systems of plants and animals wherever possible.
The presence of an introduced or exotic species in
a National Park and its effect on the native eco-
system may require historical as well as scientific
documentation prior to proposal of management plans
(United States Department of the Interior 1978).
This paper reviews the historical absence of moun-
tain goats in the Olympic Mountains and their in-
troduction by man prior to creation of Olympic
National Park in 1938. It traces the dispersal of
goats within the Park and their present distribu-
tion.
STUDY AREA
Olympic National Park is located on the Olympic
Peninsula of Washington State, along the northwest
coast of the conterminous United States. The Park
is nearly 900,000 acres (365 000 ha) in size and
occurs as two units: the interior Olympic Moun-
tains, and a Pacific coastal area fronting the
Pacific Ocean for over 50 miles (80 km). The cli-
mate is wet-maritime with a steep precipitational
gradient resulting from the abrupt rise of the
Olympic Mountains, 25 miles (40 km) inland and
over 7,900 feet (2 400 m) in elevation, and the
consequent interception of moist, westerly winds
off the Pacific Ocean. Annual precipitation ranges
from 80 to 200 inches (200 to 500 cm) along the
western slopes and valleys to a "rain shadow" of
about 20 inches (50 cm) on the northeast side of
the mountains. With such abundant moisture and a
cool climate, massive mixed-coniferous forests
occur in over 50 plant communities between sea
level and glaciated peaks.
The interior Olympic Mountains are deformed sedi-
mentary rocks of Tertiary age, ringed on three
sides by marine volcanic basalts of the Crescent
Formation. A dome-like uplifting combined with
glaciation and water erosion to shape the mountains
in steep loosely arrayed peaks and ridges, from
which 10 major watersheds radiate to the adjoining
marine coasts. The Crescent Formation is exposed
in high ridges and outcrops, which may descend
3,200 feet (1 000 m) or more and provide excellent
cliff habitat for mountain goats.
METHODS
Historical data about goat releases were compiled
from notes, correspondence, and newspaper clippings
found in Olympic National Park and Olympic National
Forest files. A major source was the Port Angeles
Evening News (hereafter PAEN; the name changed to
Daily News in 1972). Many long-time residents and
mountaineers provided early goat observations.
L. Lack of Port Angeles and C. Anderson, former
Olympic National Forest biologist, kindly made
available observations of goat distribution in the
1960's. More recently, goat sightings have been
reported annually by Park Rangers, State wildlife
agents, and backcountry visitors on provided forms.
In 1972, the senior author initiated foot and
aerial surveys to chart the distribution of the
population. Between 1972 and 1976, 35 goats were
color-marked on Mount Angeles, in the north cen-
tral mountains, to provide some known reference
about goat movements. Since 1977, an intensive
mark-and-release project has been underway to ex-
amine goat population dynamics and their effect on
the native ecosystem. The present paper, while
benefiting from this work, is primarily historical
and based on data available for the period 1925 to
1975.
RESULTS
Historical Occurrence of Mountain Goats in
Washington
Mountain goats are mountain-antelope of Eurasian
origin. Their ancestors apparently migrated to
North America over a million years ago via the
Bering Land Bridge connecting Siberia with Alaska.
Dispersal south followed the nearly continuous
chain of mountains across Western North America,
and the expansion and retreat of continental gla-
ciers (Cowan and McCrory 1970). In the last
10,000 years, as the climate warmed and continen-
tal glaciers withdrew northward, goat distribution
has progressively been restricted to high mountain
retreats in the Pacific Northwest, including
Washington.
The fossil record of mountain goats is scant.
Remains have been found in only eight widely dis-
persed localities in North America (Harrington
1971). In Washington, the only evidence is fossil
horn fragments found at Washtucna Lake (Whitman
County) in the eastern part of the State (Matthew
1902, Harrington 1971). The recent distribution
of mountain goats in Washington, on the other hand,
is well-documented by historical records and obser-
vations throughout the Cascade Range of central
Washington, from the Canadian border south to Mount
Adams (Johnson 1977). The initial record is prob-
ably by Lewis and Clark. In 1805 they reportedly
saw hides and blankets woven from mountain goat
hair by Indians along the Columbia River (Burroughs
1961).
In the isolated Olympic Mountains of northwestern
Washington, however, no early faunal remains or
ethnographic evidence of mountain goats have been
reported (C. Gustafson, Department of Anthropol-
ogy, Washington State University; and P. Amoss,
Department of Anthropology, University of Wash-
ington, pers. coram. ).
A composite list of early scientific collections
of mammals on the Olympic Peninsula was assembled
by Scheffer (1946). In 1897, a group of noted
taxonomists, including C. Merriam, E. Preble, and
V. Bailey, explored the Olympic Mountains. In
1898, a party from the Field Museum in Chicago
explored the Olympic Mountains, collecting over
500 mammal specimens (Elliot 1899). Neither of
these expeditions revealed any evidence of moun-
tain goats. In a comprehensive account of Wash-
ington mammals, Dalquest (1948) makes no mention
of goats in the Olympics. He limits their distri-
bution to the Cascade Range, with one exceptional
record from northeast Washington. Scheffer (1946)
47
OLYMPIC
NATIONAL
PARK
I
I
m ,
1
MT. ANGELES -
1
1
1
|
\i • 1910 ?
1
1
i»3i *— » y
193S • * 1
l
MT. CONSTANCE
•
.
- 1946
1
1
1941
1
1947 7 __--.'
"!_' RELEASE SITES
• 19*0 7 ■ 19(1
It k">
RELEASE & DISPERSAL
PRESENT
DISTRIBUTION
Figure 1. — Distribution of mountain goats in Olym-
pic National Park, Washington: a) release and dis-
persal, b) present distribution.
states that goats were released in the Olympic
Mountains from sources outside the State. Johnson
and Johnson (1952) in their checklist of Olympic
Peninsula mammals, treats them as an introduced
species that is slowly increasing and spreading
rather widely.
Goat Releases on the Olympic Peninsula
Late in the 19th Century, Roosevelt elk abundance
declined on the Olympic Peninsula under pressures
of settlement and market hunting. In 1909, to
protect the interior mountains and elk, President
Theodore Roosevelt proclaimed Mount Olympus a Na-
tional Monument, under the administration of the
U.S. Forest Service. By the 1920' s, elk were ap-
parently increasing under protection. County and
Federal officials received requests from State
Parks and even Alaska for elk. Local residents
and officials considered the terrain south of Port
Angeles to be ideal for mountain goats. Exchanges
were conceived to send elk calves north and moun-
tain goats south (PAEN January 2, 1925).
The first four goats were obtained from British Co-
lumbia. They were released January 1, 1925, along
the northern edge of the Olympic Mountains near
Lake Crescent (PAEN January 2, 1925; Webster 1925)
(fig. la). The record of other releases is less
clear (Webster 1932). In either 1927 or 1929, sev-
en or eight goats from Alaska were released, all
but two again near Lake Crescent. The latter were
apparently released 8 miles (13 km) to the east, in
the Elwha River valley near Baldy Ridge. A diary
kept by one of the participants, Clallam County
game warden Van Welch, is quoted later in a news-
paper account (PAEN November 4, 1947). No other
official records have been found. Welch indicated
that all goat releases after 1925 occured in 1929.
48
Later accounts, however, report that after 1925,
goats were released in 1927 or 1929. A letter
dated May 7, 1938, from Alaska Game Commissioner
F. Dufresne, recalls eight animals from Alaska
being released in 1927, two from the Chugach Moun-
tains near Anchorage and six from southeastern
Alaska. Welch's diary says six goats were from
Juneau and two from Cordova, Alaska, in 1929 (PAEN
November 4, 1947). Correspondence signed by Park
Superintendent P. Macy on October 6, 1947, also
relates that one of the eight Alaskan animals died
enroute.
Thus 11 or 12 mountain goats were released in the
Olympic Mountains between 1925 and 1929. Seven or
eitht were from Alaska, four from British Columbia.
Early accounts refer to the animals released as
"pairs," although no official records were kept of
the number of males and females.
In exchange for these goats, eight young Roosevelt
elk were shipped from Port Angeles in 1928 and es-
tablished on Afognak Island as the only elk popula-
tion in Alaska (Troyer 1960).
Goat Dispersal 1925-75
During the 1930' s, groups of up to 12 goats were
occasionally reported on the slopes of Mount Storm
King above Lake Crescent (PAEN November 4, 1937).
Observations during this period also indicate that
goats were dispersing away from the release area.
They were reported on Mount Appleton in 1929
(8 miles (13 km) south of Lake Crescent); on Mount
Angeles in 1933 (16 miles (26 km) east); at the
headwaters of the Dungeness River in 1931 (35 miles
(56 km) east); and on Mount Constance in 1935 (37
miles (60 km) east) (fig. la).
Newspaper and Ranger reports suggest that mountain
goats were established in small numbers by the
early 1940' s in the northern mountains and along
the east Park boundary. By 1947 they were observed
annually on Mount Anderson, 33 miles (53 km) south
of Lake Crescent in the interior mountains. Obser-
vations by Olympic National Forest personnel indi-
cate occasional sightings of goats south of the
Park boundary prior to 1960 and annually thereafter
(C. Anderson, B. Beckstead, pers. comm. ). Although
scant in supportive detail, these records suggest
a dispersal east and south from Lake Crescent of
up to 1 to 2 miles (2 to 3 km) per year, reaching
the east Park boundary within 10 years and the
south boundary in about 30 years.
During the 1960's and 1970' s, goat observations
were obtained from all sectors of the Olympic
Mountains, although mainly the eastern half of the
Park and lands adjoining in Olympic National For-
est (fig. lb). No goats were reported in the cen-
tral Bailey Range or around Mount Olympus until the
late 1960's. This high, central massif has the
heaviest annual snowfall (150 to 200 inches (380
to 500 cm)), which may have slowed dispersal into
the area. Few people (including scientists) visi-
ted the Bailey Range until the 1970' s, which could
also account for fewer reports. The most distant
record of goat dispersal was an animal sighted in
1973 near Grisdale, Washington, 10 miles (16 km)
below the south Park boundary and about 50 miles
(80 km) south from Lake Crescent.
CONCLUSIONS
Scheffer (1946) and Dalquest (1948) discuss the
isolation of the Olympic Mountains from the Cas-
cade Range of central Washington, and the effect
of intervening Puget Sound lowlands on the migra-
tion of certain mammals to this northwestern cor-
ner of the State. The Olympic marmot (Marmota
olympus) is apparently a preglacial alpine mammal.
Other alpine mammals found in the Cascade Range
but not in the Olympics include the bighorn sheep
(Ovis canadensis calif orniana) , the pika (Ochotona
princeps) , and the golden-mantled ground squirrel
(Citellus lateralis). Several authors have indi-
cated that lowlands and broad valleys were barriers
to mountain goat dispersal elsewhere (Klein 1965,
Cowan and McCrory 1970, Harrington 1971). It seems
reasonable to conclude that lowland and water
barriers separating the Olympic Mountains from the
Cascades were broad enough to prevent mountain
goats and certain other mammals from natively mi-
grating to the Olympics. In any case, there is
neither historic nor prehistoric evidence to sug-
gest that goats were ever native to the Olympic
Peninsula.
Introduction of mountain goats by man in this cen-
tury has successfully established a population on
the Olympic Peninsula. Their dispersal has coin-
cided with rock outcrops of the Crescent Formation,
which surrounds the inner Olympic Mountains on
three sides and, in general, has less snow than
the interior mountains. Goats are now found in
over 700 square miles (1 800 km^). An estimated
90 percent of the habitat occupied is in Olympic
National Park.
LITERATURE CITED
Burroughs, R.
1961. The natural history of the Lewis and Clark
Expedition. Mich. State Univ. Press, Lansing.
Cowan, I. McT. , and W. McCrory.
1970. Variation in the mountain goat (Oreamnos
americanus) Blainville. Jour. Mamm. 51(1):
60-73.
Dalquest, W.
1948. Mammals of Washington. Vol. 2,
Hist., Univ. Kansas Pubs., Lawrence.
Mus. Nat.
Elliott, D.
1899. Catalogue of mammals from the Olympic
Mountains, Washington. Field Mus. Chicago,
Pub. 32, Zool. Ser., 1:241-276.
49
Harrington, C.
1971. A pleistocene mountain goat from British
Columbia and comments on the dispersal history
of Oreamnos. Can. Jour. Earth Sci. 8(9):
1081-1093.
Scheffer, V.
1946. Mammals of the Olympic Peninsula, Wash-
ington. U.S. Fish and Wildl. Serv. , unpub.
manuscript, Olympic Nat. Park Library, Port
Angeles.
Johnson, M. L. , and S. Johnson.
1952. Check list of mammals of the Olympic
Peninsula. The Murrelet 33(3):32-37.
Troyer, W.
1960. The Roosevelt elk on Afognak Island,
Alaska. Jour. Wildl. Manage. 24(1) :15.
Johnson, R.
1977. Status and management of mountain goat in
Washington, p. 41-46. In Proc, First Int. Moun-
tain Goat Symp. W. Samuel and W. Macgregor,
Eds. Fish and Wildl. Br. Prov. Brit. Columbia.
Klein, D.
1965. Post-glacial distribution patterns of mam-
mals in the southern coastal regions of Alaska.
Arctic 18(l):7-20.
Matthew, W.
1902. List of the pleistocene fauna from Hay
Springs. Nebr. Bull. Amer. Mus. Hist. 16:
317-322.
U.S. Department of the Interior.
1978. National Park Service. Management Poli-
cies, Chapter IV (Natural Resource Management).
Webster, E.
1925. Status of mountain goats introduced into
the Olympic Mountains, Washington. The Murrelet
6:10.
Webster, E.
1932. Status of mountain goats introduced into
the Olympic Mountains, Washington. The Murrelet
13:25.
50
Factors Reflecting Mountain Goat Condition and Habitat Quality: A Comparison
of Sub-Populations in Olympic National Park
Victoria Stevens
ABSTRACT
An introduced population of mountain goats in Olympic National Park has provided the opportunity to study
sub-populations of different densities and conditions. During the summer of 1979, nine sub-populations
were sampled for reproductive rate, standard morphological measurements, blood values, fecal nitrogen, and
foraging behavior. Sub-population differences in hematocrit, hemoglobin, and girth correlated with dif-
ferences in reproductive rate. An increase in feeding throughout the summer unexpectedly related to a de-
crease in forage quality monitored by fecal nitrogen.
Victoria Stevens, Cooperative Park Studies Unit,
College of Forest Resources, University of Wash-
ington, Seattle, Washington 98195.
51
INTRODUCTION
Reproductive Rate
The population dynamics of an introduced ungulate
on an island have been documented on several occa-
sions (Scheffer 1957, Woodgerd 1964, Klein 1968).
In the absence of predation, populations tend to
follow a pattern of exponential growth until a re-
source becomes limiting, after which they may crash
to a few individuals. National Parks are often
relicts of natural vegetation in a sea of disturbed
habitat. The mountainous regions of Olympic Na-
tional Park in particular share with islands the
attribute of isolation. However, the island is
large enough for wi thin-island dispersal to enter
into the dynamics of an introduced species. Dis-
persal provides an outlet which may delay the
effects of over-population.
The deliberate introduction of mountain goats onto
the Olympic Peninsula more than 50 years ago has
resulted in a population that now provides an op-
portunity to observe sub-populations at different
chronological stages of development. A change in
the density of a population and the impact of den-
sity on the habitat are the major results of longer
establishment. Although the population was not
studied in earnest until 1977, 52 years after the
first introduction, it may be possible to recon-
struct the pattern of growth and dispersal and pre-
dict the future size of the population by examining
in detail differences in the sub-populations as
they exist today. Visual differences between ef-
fects of goat density on habitats in different
areas have been noted by laymen as well as scien-
tists in the Park. More precise methods, however,
for distinguishing relative impact on the habitats
and the resultant change in the condition of indi-
viduals in the resident sub-populations are desir-
able. This approach seems promising in that both
demographic and morphologic variations have been
demonstrated between conspecific ungulate popula-
tions in different physiological condition
(Nievergelt 1966, Geist 1971). The physiological
condition of a sub-population is the average
physical condition of the individuals making up
the population; and although related to habitat
quality, it may be studied directly in terms of
physiological characteristics of the individuals.
An initial step in setting up a historical con-
tinuum of sub-populations is the determination of
appropriate attributes to measure both condition
of sub-populations and quality of habitats. This
paper is the result of attempts during one summer
to determine accurate indices.
METHODS
In each of four areas, data relating to animal
condition and habitat quality were compared to the
reproductive rate. Overall reproductive rate or
fecundity was used as the indicator of relative
sub-population condition (Klein 1970, Franzmann
and LeResche 1978). Against this were compared
standard morphological measurements and blood
values for direct measures of goat condition.
Fecal nitrogen and foraging behavior were tested
as indices of comparative habitat quality.
Reproductive rate was measured as the number of
kids per 100 adult females 2 years or older. Two
year olds were included as the lowest reproductive
age parkwide since this age class demonstrated its
ability to reproduce under favorable conditions in
some sub-populations. Increased productivity may
occur by three mechanisms: reduction in age of
puberty (Caughley 1970, Markgren 1974), increases
in litter size (Zuckerman 1953, Coop 1966, Markgren
1974, Caughley 1976) and an increase in ovulation
rate (Allen and Lamming 1961, Coop 1966, Markgren
1974). An increase in reproductive rate may be
related to nutrition (forage quality and availabil-
ity) in all cases.
Standard Morphological Measurements
Some morphological measurements vary considerably
with age while others remain stable throughout the
adult life of the animal. Those that change with
overall growth may indicate the relative condition
of the sub-populations when statistical means are
compared.
We baited goats with salt and captured them with a
manually operated, rope leg snare. Their eyes were
covered to reduce stress while we took measurements
of total length, right horn length, right hind
foot, girth, and weight. Unfortunately, weight had
to be eliminated because of between population dis-
crepancies in field technique. Horn growth posi-
tively relates to nutritional status in at least
two bovids, Dall sheep (Bunnell 1978) and mountain
goats (Foster 1978). Klein (1964) demonstrated
differences in skeletal growth of deer depending
on range differences.
Blood Values
While the goat was restrained, 30 mis of blood were
taken from the jugular vein with a 19 gauge butter-
fly needle. The sample was divided among three
vacutainer tubes, one containing the anti-coagulant
EDTA (ethylene diamine tetra-acetic acid), and
stored in snow until taken to a hospital laboratory
in Port Angeles. The time between blood drawing
and analysis was 12 to 24 hours. Parameters exa-
mined were red blood cell counts (RBC), hemoglobin
(Hgb), hematocrit (Hmct), mean cell volume (MCV) ,
white blood cell counts (WBC), calcium, and phos-
phorous.
The use of blood values to assess the condition of
populations of wild mammals has increased in recent
years (Lee et al. 1977, LeResche et al. 1974, War-
ren and Kirkpatrick 1978). Considerable work has
been done relating specific parameters to condition
and to external effects such as drugs and stress
(Blankenship and Varner 1977, Franzmann and
LeResche 1978, Karns and Crichton 1978, Scanlon
1979). Franzmann and LeResche (1978) noted a sig-
nificant difference in some blood values between
moose populations of different physiological con-
dition measured by reproductive rate. They found
that anemia (indicated by low hematocrit) was one
of the best indicators of condition in moose and
not significantly influenced by stress. Hemato-
crit is the number of red cells per volume multi-
plied by the mean cell volume. It can also be
52
measured by packed cell volume. Other indicators
of good condition in moose are Hgb, calcium, and
phosphorous (Franzmann and LeResche 1978).
Fecal Nitrogen .
Between one and seven fresh fecal samples were
collected during each visit to a sample area.
These were predominately from adult females and
were frozen until the end of the season when they
were dried and ground. Nitrogen was determined by
Kjeldahl analysis with a lithium sulfate digest
(Parkinson and Allen 1975). Fecal nitrogen offers
an index of forage protein content. Price (1977)
found fecal nitrogen positively correlated to
protein intake in the hartebeest. Although this
correlation is not linear due to an increase in
metabolic fecal nitrogen (MFN) with increased dry
matter ingested (Gallup and Briggs 1948, Lancaster
1949, Hutchinson 1958) and an apparent increased
true nitrogen digestibility with increased nitro-
gen intake (Hutchinson 1958), diets can be ranked
using a fecal nitrogen index especially if dry
matter intake can be assumed to be similar between
diets.
Foraging Behavior
Foraging behavior was observed in each area to test
the hypothesis that the amount of food ingested in-
creases with an increase in forage quality; i.e.,
an increase in nitrogen and a decrease in fiber
content. It was hoped this would provide an indi-
rect means of evaluating range conditions either
between seasons or between ranges.
The hypothesis is well supported in the literature.
Ungulate feeding strategies are dominated by limi-
tations of rumen size and rate of passage. Rate
of passage and nutritive quality (protein content)
increase with decreased fiber content (Balch and
Campling 1965, McDonald et al. 1973, Milton 1979).
Therefore, to keep the rumen full, more forage is
likely to be ingested during early summer when the
quality is highest.
In 1978 more than 5,000 systematic observations of
goat activity were made in one area in the Park.
These showed that goats steadily increased the
amount of time per day spent foraging from early
June through early September (Stevens 1979). For-
aging was defined for the purposes of the 1978
observations, as either searching or feeding. The
increase in absolute foraging time coincided with
a decrease in forage quality as measured by plant
phenology and as reflected in fecal nitrogen levels
from the area. In the 1979 field season, we hoped
to explain this contradiction with the above hy-
pothesis by separating foraging time into actual
feeding time and searching time. Perhaps as the
resource becomes more patchy towards the end of
summer, the goats spend more time searching for
each bite but are actually spending less time
feeding on the lower quality forage (fig. 1).
"° 40
!■
3
o
S 30
% 20
u
4>
CL
Observed foraging
time throughout
the summi
June
July
August
/^ Theoretical
decrease in
otal feeding
time
Sept.
Figure 1. — A hypothetical explanation for the
increase in foraging time documented during the
summer of 1978 (perpendicular lines). As the
forage quality declined, the amount of searching
time increased relative- to the amount of feeding
time (diagonal lines). Theoretical search time is
the difference between total feeding time and the
theoretical feeding time.
sight, or after 1 hour when a new subject was lo-
cated and observed. The stop watch lapsed time
divided by the wrist watch lapsed time and multi-
plied by 100 indicates the percentage of total
foraging time spent actually feeding.
RESULTS
Direct Measures of Goat Condition
Reproductive rate was used as an indicator of rel-
ative condition (table 1). The best correlation
with reproductive rate among the parameters meas-
ured is with hematocrit (fig. 2). In addition,
hemoglobin and girth positively correlate with
reproductive rate when the statistical means for
adult females in each sub-population are compared.
Other measures show no correlation (tables 2 and
3). The small size of the sample in many of the
smaller, more remote areas hindeis statistical
treatment of differences between populations.
Since the largest difference occurs between Klah-
hane Ridge (the densest sub-population) and any
other area, we have lumped all the values except
those from Klahhane Ridge, including those from
areas not specifically mentioned, in tables 2
and 3.
T-tests were used to compare means from Klahhane
Ridge and means from the rest of the sub-
populations. None of the morphological measure-
ments was significantly different although girth
had a tendency to be higher in sub-populations off
Klahhane Ridge (p < .18). RBC, Hgb, and Hmct were
significantly higher in the sub-populations with
higher reproductive rates (p < .001).
Foraging adult females were timed with a wrist
watch and a stop watch. The stop watch was run-
ning only when the goat was taking a bite and
chewing; it stopped when she uras searching. The
observation was terminated by any interruption in
the foraging activity, when the goat went out of
53
Table 1 — A comparison of reproductive rates be-
tween sub-populations in Olympic National Park,
1979
Area
Reproductive Ratei'
Klahhane Ridge
Appleton Pass
Sawtooth Ridge
Royal Basin
59
60
67
75
80-,
c 6<H
0
o
k.
0)
Q. 40-
20-
KR
LJ
AP RB
_i
Productivity
Hematocrit
KR - Klahhane
AP - Appleton
RB - Royal basin
SR - Sawtooth
SR
I/Kids per 100 adult females 2 years old and
older.
Figure 2. — A comparison of reproductive rate and
hematocrit for each of four sub-populations in
Olympic National Park in 1979.
Table 2 — Mean blood values for adult females in four sub-populations of Olympic National Park, 1979
Group
N
Hematocrit
Hemoglobin
Tot
al protein
Calcium
Phosphorus
gms/dl
mg/dl
Population mean:
(male and female)
60/59
37.68
13.06
6.6
10.2
5.5
female mean
41/41
38.63
13.09
6.5
10.0
5.3
Klahhane Ridge
17/17
34.53
12.21
6.6
10.0
5.1
Appleton Pass
A/1
40.75
13.9
7.1
10.6
7.2
Sawtooth Ridge
6/5
41.83
13.87
6.5
10.0
6.4
Royal Basin
8/8
43.50
14.25
6.6
9.2
5.1
Mean of all areas
except Klahhane
24/23
41.54
13.73
6.5
10.0
5.4
Table 3 — Morphological means for adult females (4 years or older),
19791/
Group
Total
length
Hind
foot
Girth
Horn
length
Klahhane Ridge
Appleton Pass
Royal Basin
Mean of all areas
except Klahhane
143.7 (7)
145.3 (3)
148.5 (4)
34.24 (7)
36.00 (2)
34.50 (4)
99.7 (7)
102.67 (3)
106.25 (4)
22.2 (7)
22.5 (1)
20.75 (4)
143.8 (11) 34.40 (11) 103.09 (11) 21.33 (11)
^.'Number of individuals measured appears in parentheses.
54
Indirect Measures of Habitat Quality
Fecal nitrogen was used as a measure of forage
quality because of the correlation between fecal
nitrogen values- and plant phenology noted during
1977 and 1978 in the Olympic Mountains (Driver et
al. 1978). These values reflect the quality of
the vegetation if it is assumed that the phenology
of the plants dictate relative quality. When the
plants are growing most actively, the nutrient
quality of the vegetation is highest (Braun 1972).
Fecal nitrogen values from Klahhane Ridge in the
1979 samples followed the expected curve, being
highest in mid-June (fig. 3).
If forage quality can be measured between months
on Klahhane Ridge using fecal nitrogen, it should
be possible to measure differences in quality
between areas, with the same general climatic
regime, at similar time periods during the year
(assuming similar dry matter intake). Figure 3
shows the values of fecal nitrogen for the four
major sub-populations throughout the summer. The
most complete data came from Klahhane Ridge which
has been shown to have the lowest reproductive
rate. The other three sub-populations also show
decreasing forage quality throughout the summer,
but the comparison with Klahhane Ridge is not
clear. In two of the three areas, the early sum-
mer value is higher than Klahhane and the late sum-
mer value is lower indicating a more extreme change
throughout the summer. If this is a reflection of
the true forage quality, the higher reproductive
rates in these areas indicate the higher relative
importance of the early summer to the overall con-
dition of the goats. A closer look at specific
plants and parts of plants utilized between areas
throughout the summer would be helpful.
Unexpectedly, the proportion of time spent feeding
decreased with declining forage quality. Figure 4
shows how foraging behavior varied with fecal
nitrogen (forage quality) on Klahhane Ridge. On
Klahhane Ridge, where once again we have the most
complete information, the amount of time spent
feeding; i.e., the quantity of forage ingested
(assuming a constant bite volume) decreases with
increased forage quality. Ungulate feeding theory
predicts a decrease in forage consumption with a
decrease in quality because of the limitations of
the digestive system.
Appleton Pass
c
u
o
a
c
a)
o>
o
CO
u
a>
u.
June
July
August Sept.
Figure 3. — A comparison of fecal nitrogen values
from four sub-populations in Olympic National Park
during the summer of 1979: Klahhane Ridge | | ,
Royal Basin O , Appleton Pass A , and Sawtooth
Ridge • .
Figure 4.— The relationship of the percent of for-
aging time actually spent feeding as fecal nitro-
gen (protein content of the forage) increases as
demonstrated by mountain goats on Klahhane Ridge
in 1979.
E
O)
c
o
0)
0 2.202.302.40 2.502.602.702.802.903.003 10
Fecal nitrogen (percent)
55
Table 4 — Reproductive rates since 1976 on Klahhane
Ridge
Year
Reproductive ratei'
1976
1977
1978
1979
1980
78
98
25
59
72
_'Kids per 100 females 2 years and older.
DISCUSSION
The measures which showed significant differences
between sub-populations were blood values, param-
eters which reflect changes in condition most
rapidly. Of all the morphological measurements
taken, girth, while not showing statistical signif-
icance, showed the greatest difference between the
populations of different reproductive rates. Skel-
etal measurements reflect the cumulative conditions
of an individual throughout its life, whereas girth
includes fat and muscle tissue--parameters more re-
sponsive to changes in immediate condition. This
Indicates that conditions on Klahhane Ridge have
changed only recently relative to the other areas
on the Olympic Peninsula, a conclusion supported
by changes in reproductive rate on Klahhane Ridge
over the last 5 years (table 4). Overall fecundity
has gone from a high of 98 percent to 25 percent
and has been recovering for the last 2 years (Ste-
vens 1979, Stevens unpublished data). This rapid
change in reproductive rate is consistent with the
pattern observed in other introduced ungulate pop-
ulations when a resource becomes limiting.
The inconclusiveness of the use of fecal nitrogen
or foraging behavior as variables in quality com-
parisons may be a consequence of the small number
of sample points which in turn reflects the diffi-
culty in acquiring data in the more remote goat
ranges in Olympic National Park. The unexpected
trend in the foraging behavior with respect to
forage quality points out either a major fault in
the reasoning of ecologists studying ungulate feed-
ing or an omission of some critical measurement
between populations. The latter is strongly sus-
pected since current theory has been supported
repeatedly. The missing information may be the
difference in the volume of bites throughout the
summer. Bite volume was assumed to be constant;
but as forage quality decreases, it may be advan-
tageous for goats to be more selective and take
only those parts of the plant having the highest
nutritive value, thereby decreasing the size of
bites. Smaller bite volume would necessitate more
time feeding to ingest a given amount and would
have been interpreted by us to mean higher overall
forage quality.
The increased selectivity of the goats also would
result in an artificially high fecal nitrogen level
in relation to overall forage quality. This sug-
gests a mechanism for testing the changes in forage
selectivity by goats throughout the summer by com-
paring direct measures of nitrogen in plant samples
to fecal nitrogen at several intervals during the
summer. Further studies including the collection
and analysis of plants for nitrogen content are
necessary to evaluate the usefulness of foraging
behavior as an index to relative forage quality,
and therefore to the projected condition of the
sub-population.
Selected blood values appear to be useful means for
the evaluation of sub-population condition. It
would be helpful, however, to have an index such
as foraging behavior or fecal nitrogen to indicate
possible trends in population condition. Measures
which do not involve handling large ungulates are
often more efficient and less costly as management
tools.
ACKNOWLEDGMENTS
Appreciation is expressed to Lee Anne Ayres, Peter
Wimberger, Marie Morin, and Angela Alston for in-
valuable assistance in the field. This research
was funded by Olympic National Park and through the
University of Washington Cooperative Parks Study
Unit, contract no. CX-9000-7-0065.
LITERATURE CITED
Allen, D. M. , and G. E. Lamming.
1961. Nutrition and reproduction in the ewe.
J. of Ag. Sci. 56:69-79.
Balch, C. C. , and R. C. Campling.
1965. Rate of passage of digesta through the
ruminant digestive tract. In Physiology of di-
gestion in the ruminant. R. H. Dougherty, ed.
Butterworths, Washington, D.C. p. 108-123.
Blankenship, L. H. , and L. W. Varner.
1977. Factors affecting hematological values of
white-tailed deer in South Texas. Proc. Annual
Conf. S. E. Assoc. Fish and Wildl. Agencies.
31:107-115.
Braun, H. M. H.
1972. Primary production in the Serengeti:
purpose, methods and some results of research.
Annales de l'Universite d'Abidjan, Section
Biologique. 6:171-188.
Bunnell, F. L.
1978. Horn growth and population quality in
Dall sheep. J. Wildl. Manage. 42:764-775.
Caughley, G.
1970. Liberation, dispersal and distribution of
Himilayan Thar (Hemi tragus jemlahicus) in New
Zealand. New Zealand J. of Sci. 13:220-239.
Caughley, G.
1976. Wildlife management and the dynamics of
ungulate populations. In Advances in Applied
Biology. T. H. Coaker, ed. Academic Press,
New York.
56
Coop, L. E.
1966. Effect of flushing on reproductive per-
formance of ewes. J. of Ag. Sci. 67:305-323.
Driver, C. H. , V. Stevens, and D. Pike.
1978. Olympic Mountain goat — annual report.
Univ. of Washington, Cooperative Parks Study
Unit. Contract no. CX-9000-7-0065.
Foster, B. R.
1978. Horn growth and quality management for
mountain goats. Proc. of the Northern Wild
Sheep and Goat Conf. Penticton. p. 200-226.
Franzmann, A. W. , and R. E. LeResche.
1978. Alaskan moose blood studies with emphasis
on condition evaluation. J. Wildl. Manage.
42:334-351.
Gallup, W. , and H. M. Briggs.
1948. The apparent digestibility of prairie hay
of variable protein content, with some observa-
tions of fecal nitrogen excretion by steers in
relation to their dry matter intake. J. of An.
Sci. 7:110-116.
Geist, V.
1971. Mountain sheep: a study in behavior and
evolution. The Univ. of Chicago Press, 383 p.
Hutchinson, K. J.
1958. Factors governing fecal nitrogen wastage
in sheep. Aus. J. of Agric. Res. 9:508-520.
Karns, P. D. , and V. F. J. Crichton.
1978. Effects of handling and physical restraint
on blood parameters in woodland caribou. J.
Wildl. Manage. 42:904-908.
Klein, D. R.
1964. Range related differences in growth of
deer reflected in skeletal ratios. J. Mammal.
45:226-235.
LeResche, R. E. , U. S. Seal, P. D. Karns, and
A. W. Franzmann.
1974. A review of blood chemistry of moose and
other cervidae with emphasis on nutritional
assessment. Naturaliste Can. 101:263-290.
Markgren, G.
1974. Factors affecting the reproduction of
moose (Alces alces) in three different Swedish
areas. Int. Congr. of Game Biol. 11:679-690.
McDonald, P., R. A. Edwards, and J. F. D.
Greenhalgh.
1973. Animal nutrition. Oliver and Boyd,
Edinburgh.
Milton, K.
1979. Factors influencing leaf choice by howler
monkeys: a test of some hypotheses of food
selection by generalist herbivores. Am. Nat.
114:362-378.
Nievergelt, B.
1966. Der alpensteinbock (Capra ibex L. ) in
seinem Lebenstraum. Mammalia depicta. Verlag
Paul Parey, Hamburg und Berlin. 85 p.
Parkinson, J. A. , and S. E. Allen.
1975. A wet oxidation procedure suitable for
the determination of nitrogen and mineral nu-
trients in biological material. Commun. Soil
Sci. and Plant Anal. 6:1-11.
Price, M. R. S.
1977. The estimation of food intake, and its
seasonal variation, in the Hartebeest. E. Afr.
Wildl. J. 15:107-124.
Scanlon, P. F.
1979. Constraints in evaluating data from blood
of wild animals immobilized by drugs. Presented
at the 14th International Wildl. Congr. , Dublin,
Ireland. October.
Klein, D. R.
1968. The introduction, increase and crash of
reindeer on St. Matthew Island. J. Wildl.
Manage. 32:350-367.
Klein, D. R.
1970. Food selection by North American deer and
their response to over-utilization of preferred
plant species. In Animal populations in rela-
tion to their food sources. A. Watson, ed.
Brit. Ecol. Soc. Symp. 10:25-46.
Lancaster, R. J.
1949. Estimation of digestibility of grazed
pasture from faeces nitrogen. Nature
163:330-331.
Lee, J., K. Ronald, and N. A. Oritsland.
1977. Some blood values of wild polar bears.
J. Wildl. Manage. 41: 520-526.
Scheffer, V. B.
1957. The rise and fall of a reindeer herd.
Sci. Mon. 73:356-362.
Stevens, V.
1979. Mountain goat habitat utilization in
Olympic National Park. M. S. Thesis, Univ. of
Washington, Seattle.
Warren, R. J. , and R. L. Kirkpatrick.
1978. Indices of nutritional status in cotton-
tail rabbits fed controlled diets. J. Wildl.
Manage. 42:154-158.
Woodgerd, W.
1964. Population dynamics of bighorn sheep on
Wildhorse Island. J. Wildl. Manage. 28:381-390.
Zuckerman, S.
1953. Breeding seasons of mammals in captivity.
Proc. of the Zoolog. Soc. of London. 122:827-950.
57
Mother-Infant Interactions Among Free-Ranging, Non-Native, Mountain Goats
(Oreamnos americanus) in Olympic National Park
Michael Hutchins and Craig Hansen
ABSTRACT
Preliminary results are presented on nursing behavior, weaning, and retention of older offspring. Nursing
behavior was investigated on the assumption that a conflict of interest exists between mothers and off-
spring over the disposition of maternal resources during weaning. A relationship was found between the
behavior of kids during nursing attempts and the behavior used by mothers to terminate attempts. Results
suggest a progressive decrease in the overall rate of successful nursing attempts and mean duration of
suckling bouts over time, especially between the 1st and 2d month following parturition. Several cases
were identified in which females retained older offspring and continued to allow close physical contact,
including nursing. Theoretical implications of these results are discussed.
Micliael Hutchins and Craig Hansen, Department of
Psychology, University of Washington.
58
INTRODUCTION
Studies of mother-offspring relationships among
free-living animals have been accumulating stead-
ily; however, many species still lack detailed
investigation. Research of this kind is of great
interest to behavioral ecologists who strive to
understand the evolution of parental care and the
role parent-offspring relationships play in the
development of social behavior. In addition,
there has been an increased appreciation of the
potential for applying such information to the
management of wild and captive animal populations
(Geist 1971a, Cowan 1974, Lent 1974). Here we
present data on mother-offspring relationships
among a free-ranging, non-native population of
mountain goats (Oreamnos americanus) in Olympic
National Park, Washington. This preliminary anal-
ysis is based on data collected during a compre-
hensive, on-going study of mountain goat social
biology and is far from complete in either its
scope or conclusions. Three topics are discussed:
nursing behavior, weaning, and the retention of
older offspring.
The mountain goat is a large herbivore that typi-
cally inhabits rock outcrops and alpine meadows
lying at or above timberline (Walker 1975, Rideout
1975). The mating system is polygynous; adult
males and females live apart for most of the year,
associating consistently only during the rut which
occurs in late November and early December (DeBock
1970, Geist 1965). Gestation lasts about 178 days,
and kids are born in late May or early June (Ride-
out 1974). Adult females typically isolate them-
selves at the time of birth, but on summer range
they congregate with juveniles of both sexes and
kids in loosely organized bands where a variety of
social interactions has been observed (Brandborg
1955, Holroyd 1967, DeBock 1970). The natural dis-
tribution of 0_. americanus is restricted to the
mountains of western North America (Cowan and
McCrory 1970). Because of the relative inaccessi-
bility and rugged nature of its habitat, it remains
one of the least studied of the continent's big
game animals.
STUDY AREA AND POPULATION
Olympic National Park is a 3 600-km2 island of
wilderness occupying a majority of the Olympic
Peninsula in northwestern Washington State. The
Olympic Mountains are uniquely isolated, being
surrounded on three sides by salt water and by
lowlands on the fourth. The mountains, which form
the core of the Park, are dominated by 2 428-m-high
Mount Olympus and represent a complex concentra-
tion of peaks, ridges, and valleys.
The region's climate is maritime and influenced
heavily by the easterly flow of weather off the
Pacific Ocean. Summers tend to be relatively cool
and dry, and winters are relatively mild and wet.
Precipitation patterns are influenced by topog-
raphy; northern and eastern portions of the Park
receive considerably less rainfall than do coastal
areas. At high elevations, winter precipitation
is usually in the form of snow.
Research efforts were concentrated on Klahhane .
Ridge which is located in the northeastern portion
of the Park. This 1 829-m-high ridge is bounded
by 1 950-m Mt. Angeles on the west and by 1 889-m
Rocky Peak on the east. The area is geologically
complex, being composed primarily of pillow basalt
and volcanic breccia (Tabor 1975). It has been
described as "excellent" mountain goat habitat,
with rugged, steep terrain above timberline and
sufficient vegetation for forage (Olmstead 1976).
The ridgetop is narrow with the major vegetational
habitats being distributed on north- and south-
facing slopes. Vegetation of the area is dominated
by stands of Krummholz and meadows composed of
hardy alpine perennials.
No natural mineral licks occur within the study
area, but the remains of a large artificial lick
still exist. Park Service officials discontinued
the practice of provisioning salt several years
ago; but goats still use the area extensively in
the late spring and early summer, apparently seek-
ing minerals that have leached in the soil.
Several potential predators and competitors of
goats occur in the Olympic Mountains. Although no
conclusive evidence of predation has been discov-
ered in the study area, possible predators include:
the cougar (Felis concolor) , bobcat (Lynx rufus) ,
coyote (Canis latrans) , black bear (Ursus ameri-
canus) , and golden eagle (Aquila chrysaetos). Po-
tential competitors include the abundant Columbian
black-tailed deer (Odocoileus hemionus), Olympic
marmot (Marmota olympus), mountain beaver (Aplodon-
tia rufa ) , and snowshoe hare (Lepus americanus).
Mountain goats are not indigenous to the Olympic
Mountains. The present inhabitants of the Park
are descended from 11 or 12 individuals, which
were released between 1925 and 1929 (see Moorhead
and Stevens in this report). The animals were
obtained from the Canadian Rocky Mountains near
Banff, Alberta, and from two separate locations in
Alaska. The absence of native mountain goats in
the Olympic Mountains appears to be due to geo-
graphic isolation. The species is native to the
Cascade Range, less than 160 km to the east. The
Olympic Peninsula currently supports a population
estimated to be close to 700 (Driver et al. 1979),
and Klahhane Ridge has the highest concentration
in the region. As many as 165 animals use the
13-km2 summer range.
METHODS
Behavioral data were collected during 30-minute
sample periods using a continuous sampling tech-
nique (Altmann 1974). Observations focused on
identifiable mother-offspring pairs. Focal pairs
were chosen at random, with the exception of fe-
males associating with yearlings or 2-year-olds,
which were selected for their particular interest
to the study.
To facilitate data collection, a behavioral inven-
tory (ethogram) was formulated and a two- or three-
letter phonic code was devised to represent each
behavioral category (Appendix I). Time of occur-
rence (within each minute) and the type of behavior
exhibited were recorded for each behavior of inter-
est occurring during the sample period. Identity
59
Q AM OBSERVATIONS
Q PM OBSERVATIONS
30n
o24
F. 18
L
L,
»!
nil
4 5 II 14 15 18 19 20 21 22 25 27 31 32 33 37 39 49 102 110 135
FOCAL MOTHER -OFFSPRING PAIR
Figure 1. — Distribution of observation time by
timeblock for each designated focal pair in 1979.
of the actor and recipient were recorded for each
social behavior to preserve the directionality of
the interaction. Uncoded behaviors of interest or
behaviors occurring outside the sample period were
recorded in the form of qualitative or "ad libitum"
(Altmann 1974) notes.
Several rules were formulated to disperse samples
as evenly as possible between subject pairs and to
provide some measure of randomization. While in
the field, observers made daily trips through the
study area in an effort to locate focal pairs. If
two or more pairs were encountered simultaneously,
selection for observation was based on the number
of previous samples collected within the timeblock.
(AM: OOOOh to 1200h or PM: 1200h to 2400h). The
pair having the fewest samples in that timeblock
was consistently given preference, although a max-
imum of two samples/timeblock per day (one hour)
was allowed on any one pair in 1978, and a maximum
of four samples/timeblock per day on any one pair
in 1979.
To minimize potential influence of observers on be-
havior, observations were conducted from a minimum
distance of 10 m, except when special circumstances
necessitated closer approach. This was sometimes
necessary when animals moved through terrain that
made following or observing difficult. Every ef-
fort was made not to disturb the animals, but no
special attempts at concealment were deemed neces-
sary since the animals were accustomed to people
and relatively tolerant of their presence.
Over 100 animals
individual identi
plished with the
placed in the vie
were marked with
inserted into the
applicator. The
easily from great
yellow in colorat
numerals on both
left ear and male
the age of adult
the number of ann
of the horns (Bra
have been captured and marked for
fication. Capture was accom-
use of hand-held rope foot-snares
inity of the salt lick. Animals
Y-Tex Cattle ear tags. Tags were
pinnae with a specially designed
8.4-cm-long plastic tags are seen
distances, as they are bright
ion and have 2. 54-cra-high black
sides. Females were tagged in the
s in the right ear. Estimates of
females were obtained by counting
ual growth rings on the surface
ndborg 1955).
RESULTS
This preliminary analysis is based on 54.5 hours
of quantitative data collected on 9 mother-kid and
7 mother-yearling pairs from June 27 to August 23,
1978, and on 312 hours collected on 17 mother-kid,
2 mother-yearling, and 2 mother-2-year-old pairs
from June 12 to August 24, 1979. The distribution
of observation time for each focal pair during 1979
is presented by timeblock in figure 1 and indicates
that observations were distributed relatively
evenly between timeblocks, and between most focal
pairs.
Nursing Behavior
The following analysis is based on a total of 461
nursing attempts recorded during continuous sam-
pling in 1979, and is restricted to observations
of mother-kid pairs. Offspring attempted to nurse
when the mother stopped her forward progress during
travel or foraging, when she arose from a reclining
position, when she was preoccupied with another
activity (e.g., agonistic interactions with other
conspecif ics, licking salt, or urination), and upon
reunion of the pair after a brief separation. In
some instances, offspring appeared to stimulate
their reclining mothers to stand so that they could
nurse. They accomplished this by various methods,
usually by climbing on or over the mother, or by
nuzzling the udder region. Nursing was not
observed while the mother was lying down.
When offspring approached the udder during a
nursing attempt; they did so in a variety of ori-
entations with respect to the mother's body. We
recognized four basic modes of approach: Side (S),
Front (FR), Rear (R), and Run Around (RA) (Appen-
dix I). A typical S approach is depicted in fig-
ure 2. In the overwhelming majority of cases,
offspring approached the udder in a short, rapid
run, although running approaches were never used
during R-oriented attempts. During 1979, a major-
ity of attempts (81.4 percent) were initiated from
the mother's side. R attempts comprised 10.3 per-
cent of the total; RA and FR approaches accounted
for 6.5 and 1.9 percent. Besides being the most
frequently utilized of modes, S approaches also had
the highest success rate; 44 percent were success-
ful (3 seconds or longer) compared with 27 percent
of RA approaches, 22 percent of FR approaches, and
21 percent of R approaches.
While nursing, kids frequently exhibited "bunting"
behavior — a vigorous butting of the mother's udder.
An average of 4.7 bunts was delivered to the udder
during each successful attempt.
All nursing attempts recorded during continuous
sampling were terminated by mothers and not by
kids. Four primary modes of termination were ob-
served: Step-Over (SO), Walk-Away (WA), Rear Leg
Stamp (RLS), and Jump- Away (JA). All other forms
60
of termination were subsumed under an "other"
category (Appendix I) and described in qualitative
notes. The most common method utilized by females
during 1979 was SO, which accounted for 63.7 per-
cent of all terminations. Next most frequent was
WA (27.6 percent), followed by RLS (A. 3 percent)
and JA (2.0 percent). The use of aggression to
terminate nursing was very infrequent, represent-
ing only 1.5 percent of the total. This usually
consisted of a low intensity horn present, horn
swipe, or butt (Appendix I). Only four cases of
maternal aggression involved any physical contact
with the offspring. Only a single ad libitum ob-
servation was made in which an offspring appeared
to terminate a bout on its own initiative, this
occurred after a 23-second nursing bout.
In 1979 we found an apparent relationship between
the approach mode used by kids and the termination
mode used by mothers (fig. 3). When kids approached
the udder from the side, either directly or in a
RA, females terminated most frequently with a SO
or WA. This was also true of FR approaches. R
approaches were terminated most frequently with a
WA or RLS. RLS was never used to terminate a FR
or RA approach, and was used very infrequently
during S approaches. JA was associated only with
R and S approaches; maternal aggression was seen
only during S approaches.
Retention of Older Offspring
During 1978 and 1979, we documented 22 separate
cases in which identifiable females consistently
associated with yearlings or 2-year-olds (table 1).
The mean estimated age of females associating with
yearlings (N=22) was 5.7 years; that for females
associating with 2-year-olds (N=2) was 6.0 years.
Ages ranged from 3 to 10 years. Among yearlings,
9 of 25 (37.5 percent) were males, 8 of 24 (33.3
percent) were females, and 7 (29.2 percent) were
not sexed. Among 2-year-olds, one was male and one
was female.
We are convinced that these relationships represent
cases of offspring retention in females that either
lost their young of the year or did not conceive.
For 4 of the 20 cases (20 percent) in 1978, direct
evidence was available from tagging. Animals 51,
52, 53, and 54 were born in the spring of 1977 and
tagged in August of that year while still associat-
ing with their mothers. During 1978, all continued
to associate with their mothers through late August
when observations were terminated.
The conclusion is also supported by a closer exami-
nation of behavior. While adult females generally
tended to be antagonistic toward other conspecif-
ics, all instances of retention were characterized
by relatively frequent "friendly" contact, includ-
ing nursing. During 1978, 10 attempts to nurse by
five different yearlings were recorded during con-
tinuous sampling and an additional 11 attempts were
observed during ad libitum observations. Including
ad libitum observations, nursing was seen among 10
of 20 identifiable mother-yearling pairs. All at-
tempts were directed toward the female with which
that yearling had shown consistent association.
Figure 2. — Typical side-oriented nursing attempt.
Figure 3. — Approach mode of kid by termination mode
of the mother.
80
RUN
AHUUNU
60
-
(N=30)
40
-
20
80
REAR
< 60
-
V.H— 4/ )
O 40
o
t-
^80
O
cr 60
F RONT
(N=9)
°- 40
20
1 1
80
SI DE
60
40
20
i 1 i 1
SO WA RLS J A
TERMINATION MODE
61
Table 1 — Instances of offspring retention during 1978 and 1979
Female Focal
Year tag subject
number
Estimated age
at time of
retention
Offspring
retained
Offspring produced
by year
1977 1978 1979
Retained
offspring
number (sex)i'
1/
Date offspring
tagged
1978£'
1979
J 7
*
13
1/ 17
18
3/ 19
21
*
23
*
27
28
29
*
32
33
*
35
36
37
*
38
39
106
*
109
(1-side)
Green tag
5
*
14
*
18
*
27
*
4
Yearling
1
0
0
44(M)
6-78
4
Yearling
1
0
1
52(F)
8-77
7
Yearling
2
0
1
128(F), UM(?)
7-78
5
Yearling
1
0
0
UM(F)
7
Yearling
2
0
0
UM(?), UM(M)
7
Yearling
1
0
*/l
107 (M)
6-78
7
Yearling
1
0
1
108(M)
6-78
5
Yearling
1
0
0
78(M)
7-79
4
Yearling
1
0
1
UM(?)
10
Yearling
1
0
1
115(F)
6-78
3
Yearling
1
0
1
UM(?)
8-77
7
Yearling
1
0
1
51(M)
8-77
5
Yearling
1
0
1
UM(?)
4
Yearling
1
0
1
54(F)
6-78
6
Yearling
1
0
1
43(F)
6-78
6
Yearling
1
0
1
UM(?)
4-5
Yearling
1
5/0
1
UM(M)
9
Yearling
1
0
0
117(M)
6-78
5
Yearling
1
0
1
UM(?)
Yearling
53(M)
8
Yearling
1
1
0
UM(F)
5
Yearling
1
1
0
UM(F)
6
2-year-old
1
0
0
UM(F)
6
2-year-old
1
0
0
78(M)
8-77
6-79
1/
UM indicates an unmarked animal.
—'Includes only identifiable females.
3/Early in 1978, both 17 and 19 associated with two yearlings that were presumably their twins from the preceding
year. 128 continued the association with 17 throughout the summer, while the unmarked yearling did not. Similarly,
only one of 19 's yearlings was seen with her in August.
_L'Kid born to 21 in 1979 disappeared in late July and was presumed dead.
_'Kid born to 39 in 1978 disappeared in early June and was presumed dead.
Although the sample is admittedly small, the over-
all rate of successful nursing attempts/hr observed
was higher in kids than retained yearlings (0.42
vs. 0.09/hr). The mean duration of a successful
bout, however, was slightly higher for retained
yearlings than for kids (6.5 vs. 5.8 s). Only 2
(18.2 percent) of all the 11 bouts by yearlings
recorded ad libitum were "successful," one lasting
for 3 seconds and the other for 4 seconds. Kids
had a much higher rate of attempt/hour of observa-
tion (5.9 times) than did yearlings. During 1979,
no instances of nursing were observed in the two
mother-yearling pairs; but this may be a result of
small sample size; both pairs disappeared early in
the field season. One unsuccessful nursing attempt
was observed between female 18 and her unmarked fe-
male 2-year-old; however, this relationship dete-
riorated after the 2-year-old was radio-collared.
Radio-collars may interfere with the ability of
mothers to recognize their offspring. We have ob-
served radio-collared kids threatened by their own
mothers for up to 3 days after their release.
Tagging, however, appears to have no dlscernable
affect on mother-kid relationships. We did observe
six instances of nursing between female 27 and her
retained 2-year-old male; two of these occurred
during 24 hours of continuous sampling; the other
four instances were recorded ad libitum. Two of
the attempts were successful, one lasting 40 sec-
onds and the other 12 seconds.
A variety of other social interactions occurred
between females and retained offspring. Many re-
sembled those observed between mothers and kids.
For example, females allowed retained offspring to
maintain a close relationship while using the salt
lick. This often involved close physical contact,
such as leaning against one another, or licking
salt face to face in the same location. Instances
of close physical contact at the lick were recorded
for all females having retained offspring in 1979,
and for five of seven focal mother-yearling pairs
in 1978. No other individuals — regardless of age
class — were tolerated in this manner. This obser-
vation is particularly significant since the inci-
dence of overt aggression in the vicinity of this
clumped, defensible resource was comparatively high.
62
Retained offspring continued to bed in close asso-
ciation with their mothers and to exhibit a strong
following response. In two instances during 1979,
retained offspring produced loud "bleating" vocal-
izations when separated from their mothers. The
behavior closely resembled that exhibited by
mother-kid pairs. For example, on June 23, female
18 and her retained 2-year-old female became
visually separated from each other in fog. The
2-year-old vocalized three times in 4 minutes and
appeared to be searching for her mother. Female
18 responded to her offspring's vocalizations by
vocalizing herself; both were vocalizing just prior
to their reunion and both ceased calling upon being
reunited.
Eight instances of allogrooming (Appendix I) were
recorded between three of nine mother-yearling
pairs during continuous sampling in 1978; and al-
though no instances were observed among mother-
2-year-old pairs in 1979, the two mother-yearling
pairs exhibited 17 instances during 10 hours of
observation. Twenty-one of 25 total grooming
bouts (84 percent) observed between mother-yearling
pairs involved offspring grooming their mothers.
In 1979, the unmarked female yearling of female 14
groomed her mother extensively. Nearly all of the
15 instances recorded (92 percent) were directed
toward the mother's right eye which was seriously
injured. No instances of allogrooming were ob-
served between individuals who were not mother and
offspring.
Females were never observed to actively defend
older offspring from the threats of other conspec-
iflcs, but offspring occasionally moved into close
contact with their mothers when approached by a
dominant individual. During 1979, at least eight
instances were observed in which retained year-
lings or 2-year-olds moved into physical contact
with their mothers following the approach or threat
by another conspecific. This also was true for
kids; mothers tended to defend their own personal
space, and when their kids remained within this
"sphere of influence," they were protected by her
antagonism toward intruders. When they ventured
away, however, they could be threatened without
maternal retaliation.
Weaning
During 1979, females regularly terminated their
offsprings' attempts to nurse 2 weeks after birth.
Since systematic data were not collected until
June 12, there was no way of determining the exact
age at which mothers began to reject attempts on a
regular basis; nor is it known if there is any in-
dividual variation in this regard. Assuming that
most births occurred during the last week of May
or 1st week of June, qualitative observations of
nursing behavior suggested that terminations began
very early, perhaps during the 1st week of life.
Some females rejected nursing attempts as early as
May 27 (M. Morin, pers. comm. ).
Table 2 — Nursing data by month for mother-kid pairs
during 1979
June
July
August
N of mother-kid pairs
17
17
y i6
Hours of observation
70.5
112
103.5
Total nursing attempts
observed
183
150
128
N of "successful" attempts
(13s)
59
38
37
Rate of "successful" attempt/
hour of observation
.84
.34
.36
Mean duration of "successful"
attempts
14.0
9.7
8.3
Standard deviation
9.4
4.1
3.4
Range
40
15
13
i/Kid of female 21 lost or deceased in late July.
Although detailed statistical analyses have not
been undertaken, the results suggest a trend to-
ward lower nursing durations from June to August
(table 2). In the 17 mother-kid pairs observed in
1979, the rate of successful attempts/hour of ob-
servation and the mean duration of successful at-
tempts were comparatively higher in June than
either July or August and the maximum duration re-
corded during June was more than two times longer
than in July or August.
DISCUSSION
Nursing Behavior
Maternal-offspring interactions involved in nurs-
ing attempts and terminations can be viewed as an
indicator of weaning. Trivers (1974) proposed the
concept of "parent-offspring conflict" as an expla-
nation for the weaning phenomenon based on modern
evolutionary theory. The theory predicts a con-
flict of interest between mother and offspring over
the disposition of maternal resources (in this case
milk), with the offspring wanting more than the
mother has selected to give. Viewed from this per-
spective, interactions between mothers and offspring
take on new meaning. In an attempt to exploit ma-
ternal resources as best they can, offspring are ex-
pected to employ behavioral "strategies," while
mothers are expected to display appropriate "counter-
strategies" that prevent over-exploitation. When we
use the term "strategy," we do not imply that animals
make conscious decisions. We simply refer to behav-
ioral traits which have been favored by natural
selection over other alternatives.
Looking at mother-offspring interactions from that
viewpoint, offspring should employ behavior that
results in more nursing time. Alternatively, moth-
ers should limit unrestricted access to the udder
but, at the same time, minimize risk of injury to
offspring. Energy expended on behavior associated
with nursing reduces the amount available for other
activities, such as those involved in maintenance
or reproduction. The results seem consistent with
this interpretation.
63
Methods of termination employed by females seem to
be situation-appropriate. When offspring approach
the udder from the S or FR, they position them-
selves below the mother's abdomen and directly an-
terior to her rear legs. In such a position they
impede her ability to move forward. To terminate
an attempt, some forward movement appears necessary
to dislodge the kid from the nipple. SO's may re-
duce the possibility that a female steps on and
injures her offspring when it is positioned beneath
her body and, at the same time, allow her to move
forward and terminate the attempt.
When her kid is positioned beneath her body, a
female may also employ WA; however, as Brandborg
(1955) and Chadwick (1975) have noted, those ter-
minations often result in the kid being "bowled
over. " This may explain why WA is seen so infre-
quently when compared with the SO, even though it
involves less movement and is therefore likely to
be less energetically expensive. Forcibly knock-
ing a kid from the udder may increase the risk of
injury, particularly when nursing occurs on cliff
ledges, steep scree slopes, or other precarious
locations. That might also explain why WA was the
most common method of termination in response to R
approaches. Because the kid is behind its mother,
it cannot impede her forward progress and also is
in no risk of being stepped on or falling if she
does move. The few SO's that occurred during R
approaches happened in those rare instances when
the kid was able to position itself beneath its
mother.
RLS was seen most frequently during R approaches,
which suggested that it was relatively ineffective
once the kid had ahold of a nipple. It may be ef-
fective against R approaches because kids appeared
to experience some difficulty in grasping a nipple
from that orientation. DeBock (1970) thought that
R-oriented attempts may be difficult because of the
forward positioning of the teats and relatively
small size of the mountain goat udder.
The infrequent expression of aggression during
nursing terminations also suggested that risk of
injury to offspring was a major determinant of
maternal behavior. Aggression was used only to
terminate side-oriented attempts, the most success-
ful mode for offspring. Females apparently employ
aggression only when probability of successful
exploitation is high. Additionally, the fact that
JA or running away from offspring occurred so in-
frequently may be due to the higher energetic cost
as compared to other methods of termination. Both
involve a good deal more movement than other
methods.
Offspring also appeared to exhibit behavior which
can be interpreted as advantageous. In approach-
ing the udder from the S or FR, offspring position
themselves so that it is difficult for mothers to
terminate nursing attempts. Additionally, the S
approach is the most direct route to the udder.
Although RA and S approaches involved the same ba-
sic orientation during nursing, S approaches were
considerably more successful. This difference may
be due to a higher probability of early maternal
detection. In RA and FR attempts, kids passed
within their mother's primary visual field before
making contact with the udder. The R approach does
not share this disadvantage and the S approach is
probably less subject to early detection than the
RA or FR modes. That may explain why R attempts
were never preceded by a running approach; kids may
not have to move in rapidly since the probability
of early detection is low.
R-oriented attempts were rare, which attests to the
difficulty in nursing from this position. Anatom-
ical positioning of the udder and teats, and ease
of termination by the mother may render it rela-
tively disadvantageous. It is the only method of
approach in which the apparently energetically in-
expensive RLS seems to be a highly effective form
of termination.
The RA approach may have an additional advantage
that could account for its higher rate of success
when compared with the R or FR modes. This mode
appeared situation-specific, being exhibited when
the mother was moving, either in foraging or trav-
el. Among others, Geist (1971b) viewed RA as a
behavioral strategy used by young ungulates to
restrict their mother's movement so that they can
nurse. Nursing attempts directed at the mother
when she is moving are unlikely to be successful,
since it makes grasping the nipple a difficult
task.
Bunting behavior is a universal behavioral pattern
in young ungulates (Lent 1974) and can also be
viewed as a strategy. Vigorous stimulation of the
udder may result in the release of more milk and/or
stimulate further milk production.
The situations in which kids try to suckle also
appear to be advantageous. Chadwick (1975) noted
that kids "exploited minor events" to gain access
to the udder. In the present study, attempts
tended to be made when mothers stopped forward
movement, stood up from a reclining position, were
preoccupied with other activities, and upon reunion
of the pair. Kids may increase the probability of
gaining access to a teat by attempting to nurse
when mothers are preoccupied with some other activ-
ity. For example, an increase in attempts to nurse
while the mother was urinating was noted toward the
end of the summer when the rate of success was com-
paratively low. It is possible that an element of
surprise may be important, and that the probability
of early detection is lower when the mother is en-
gaged in some other activity. Once the kid begins
nursing, it can be very difficult to dislodge. We
noticed kids continuing to hold onto the nipple af-
ter mothers had stepped over them and were walking
forward in an attempt to pull them off. Older,
presumably stronger kids sometimes were dragged a
considerable distance before letting go. Kids at
times appear to stimulate their mothers to rise so
that they can nurse which supports the notion that
offspring are not passive by-standers in the wean-
ing process. Similar behavior has been noted in
other ungulates, such as domestic sheep (Ewbank
1967).
64
Retention of Older Offspring
Little is known about the breakdown of social bonds
between mother and offspring in ungulates (Lent
1974). Brandborg (1955) and Hanson (1950) believed
that mountain goat young remained with their moth-
ers until April or May of the year following partu-
rition and that they are driven off by the female
just prior to the birth of a new offspring. Hol-
royd (1967) states that yearlings may rejoin their
mothers after parturition has taken place, but that
they are kept at a distance and "treated no differ-
ently from any other intruder if the welfare of the
newborn kid is threatened in any way." Chadwick
(1975) noted that dissociation of females and kids
occurred in April. He believed the break-up to be
primarily passive, although some low intensity
maternal aggression appeared to be involved. As
maternal interest waned, young began to follow
other females and sub-adults. He felt that the
only permanent relationship among goats was the
association of females and offspring from birth to
10 to 11 months. He does describe two separate in-
stances, however, in which females "adopted" year-
lings and "treated them like kids in most respects"
(p. 146). Those yearlings were allowed to travel
and bed in close association with their female
companions. DeBock (1970) stated that the bond
between mountain goat mothers and their offspring
was broken at the birth of a new infant; however,
he did observe three instances of nursing by year-
lings which led him to suggest that barren females
may occasionally allow them to remain in close
proximity. Unfortunately, DeBock and Chadwick did
not have marked animals, and their conclusions con-
cerning length or permanency of association were
therefore limited to speculation.
We suggest that the retention of older offspring
in ungulates may represent an evolved reproductive
strategy (Hutchins 1980). Associations between
yearlings and barren females have been reported in
a wide variety of species and may be much more com-
mon than now realized (Lent 1974, Hutchins 1980).
Chadwick' s (1975) suggestion that close relation-
ships between females and yearlings represent cases
of adoption is probably incorrect, since this im-
plies that the pair is genetically unrelated.
Several studies have shown a strong correlation be-
tween habitat quality, population density, physical
condition of females, and reproductive success in
ungulates (Geist 1971b, Sinclair 1977, Wilson and
Hirst 1977). When habitat quality becomes poor, or
competition intense such that reproduction is cur-
tailed, then females may be able to increase their
previous offspring's probability of survival and/or
competitive ability by continuing their associa-
tion. For example, retained offspring appear to
share in their mother's higher dominance status
while utilizing limited resources, such as the salt
lick. Such an advantage, may extend into the win-
ter, when deep snow covers vegetational habitats,
and there is competition for patchily distributed
forage. Petocz (1973) has noted an increase in
aggression among mountain goats under conditions
of deep snow, and competition for forage has been
noted in other ungulates with similar winter range
conditions (Denniston 1956). Older offspring are
also allowed to nurse occasionally, thus possibly
having access to an energetically rich food sup-
plement that is not available to unattached mem-
bers of the same age class (see discussion of
weaning, this paper). If mothers were not lac-
tating, they appeared to incur little cost in
continuing the association and may have received
some benefit. We have noted offspring grooming
serious injuries sustained by the mother. This
may help to promote healing and therefore decrease
the chances of infection as, for example, in ro-
dents (Li et al. 1980). Females may also gain
some competitive advantage. For example, female
27 and her retained male 2-year-old (78) were
observed to approach the salt lick in tandem,
simultaneously threatening and displacing other
conspecif ics.
The hypothesis that retention is likely to be as-
sociated with adverse environmental conditions and
reduced reproductive success is indirectly sup-
ported through data on the Klahhane Ridge females
(Driver, Stevens, and Pike 1979). In 1977, the
estimated ratio of kids to adult females was
97:100, but in 1978 this figure dropped dramati-
cally to 28:100. Low reproductive success was
associated with a high incidence of retention in
1978. It is evident from table 1 that all known
cases of retention took place during years in which
females did not reproduce. Additional evidence
indicating an environmental or density dependent
influence on reproduction in 1978 comes from data
on the timing of birth and incidence of dispersal.
Kids were born 2 weeks later on the average in 1978
than in 1977 or 1979. Additionally, the incidence
of dispersal was considerably higher in 1978 than
in 1977 or 1979, it is estimated that 17 percent of
the tagged population left Klahhane Ridge during
1978 (Driver et al. 1979). It has been suggested
that a comparatively low snowpack during the pre-
ceding 2 years may have had a negative effect on
forage quality and that this combined with the high
population density acted together to reduce repro-
ductive capability.
Weaning
In comparison with studies conducted on parturi-
tion and bond formation, few exist on the process
of "weaning" in ungulates (Lent 1974); and in many
cases, the term has been used rather loosely. Lent
(1974) has noted that most authors do not make a
strong distinction between an offspring's transi-
tion from milk to forage and the breakdown of so-
cial bonds between mother and offspring. The two
processes are not always synonomous (Lent 1974).
Here we restrict our discussion to the former,
which we refer to as "nutritional weaning" to avoid
confusion.
65
Our findings concerning nutritional weaning are in
general agreement with other reports. Brandborg
(1955) observed kids foraging and ruminating when
less than a week old; at 6 weeks, they were feeding
regularly on vegetation. Chadwick (1975) found
that females rejected nursing attempts in the 2d
week after birth. He felt that kids were weaned
between 3 and 4 weeks of age, basing his conclu-
sion on the observation that suckling durations had
"stabilized at less than 15 seconds'* and nursing
attempts were "sporadic and usually rejected." He
did not provide quantitative data on rejection fre-
quency. Furthermore, he observed nursing until
mid-November and kids attempted to nurse as late
as January. DeBock (1970) thought that weaning
occurred at 4 months of age and noted that the mean
duration of nursing bouts stabilized at 15 seconds
once the young were 6 weeks old. He also found a
progressive decrease in the frequency and duration
of bouts over time. His results are questionable,
however, since he did not have marked animals and
did not collect data systematically. Ad libitum
observations do not control for observer bias or
for differences among particular mother-infant
pairs.
Our results suggest that a progressive decrease in
nursing occurs over time, but we still hesitate to
classify kids as weaned even by the end of August.
Of course, the problem of whether an offspring is
weaned or not is chiefly one of semantics. Nutri-
tional weaning could be considered to be complete
when (1) an offspring has shifted its nutritional
dependence from mother's milk to forage, or (2)
the offspring fails to receive any nourishment — no
matter how seemingly trivial — in the form of moth-
er's milk. If the former definition is employed,
then we might follow the reasoning of Chadwick
(1975) and consider kids weaned at 4 weeks when
there is a dramatic drop in the frequency and dura-
tion of successful nursing attempts. If, however,
the latter definition is employed, the possibility
exists that complete weaning may not occur for up
to 2 years. Unfortunately, we were unable to de-
termine if females with retained older offspring
were lactating. If those females were tolerating
older offspring because they lost young-of-the-
year, then they may have been physiologically ca-
pable of producing milk. Alternatively, nursing
by yearlings may simply represent a form of non-
nutritive "comfort suckling" as seen in many young
mammals, as for rats (Kenny et al. 1979) or ponies
(Tyler 1972). Considering the enormous energetic
costs involved in lactation (Hanwell et al. 1977),
we would predict nursing by yearlings and 2-year-
olds to be non-nutritive and that prolonged suck-
ling may function only in the maintenance of social
bonds.
The meaning of the phrase "nutritional dependence"
has important implications for a definition of
nutritional weaning. While even 1- to 2-week-old
kids do not appear to be completely dependent on
mother's milk as a source of nutrition, it is un-
doubtedly important, particularly when forage
quality is poor. Milk is considerably higher in
energetic value than forage, and even small amounts
may provide some nutritional advantage. Brandborg
(1955) has noted that young-of-the-year are the age
class most susceptible to mortality and suggests
that failure to over-winter is the primary cause of
death. This would suggest that a kid's physical
condition just prior to the onset of inclement
weather may affect its ability to survive. Among
bighorn sheep (Ovis canadensis) , larger lambs are
more likely to survive the rigors of winter than
are smaller lambs (Geist 1971b). Viewed from this
perspective, a few extra suckling bouts may make
the difference between survival and death.
The point to be made here is that the process of
nutritional weaning needs to be considered on the
level of the individual mother-infant pair. Con-
siderations of weaning on a populational level do
not take into account variations in the ability of
mothers to produce milk or in the need of partic-
ular offspring for this resource. Efforts to de-
fine weaning as a populational characteristic are
gross generalizations and provide only superficial
insight into an issue of immense complexity.
ACKNOWLEDGMENTS
Funding for this research was provided by grants
from the Theodore Roosevelt Memorial Fund of the
American Museum of Natural History and Sigma Xi ,
the Scientific Research Society of North America.
We would like to extend our appreciation to the
staff of Olympic National Park, especially Mr.
Bruce Moorhead, Research Biologist, and Mr. Roger
Contor, Superintendent, for their interest and
assistance. Thanks are also due to Mr. Raymond
Rasker and Ms. Margaret Davis who served as field
assistants during 1979, and to Mr. Kim Agrimson
for figure 2. Much of the data were collected in
collaboration with Ms. Victoria Stevens who has
been studying the population biology of the Olym-
pic goats since 1977; her friendship and assist-
ance are greatly appreciated.
LITERATURE CITED
Altmann, J.
1974. Observational study of behaviour.
49:227-267.
Be ha v.
Brandborg, S. M.
1955. Life history and management of the moun-
tain goat in Idaho. Idaho Dept. Fish Game Wildl.
Bull. 2:1-142.
Chadwick, D. H.
1975. Mountain goat ecology-logging relation-
ships in the Bunker Creek Drainage of Western
Montana. Unpubl. M.S. thesis, Univ. Mont.,
Missoula.
Cowan, I. McT.
1974. Management implications of behaviour in
the large herbivorous mammals, p. 921-934. I_n
V. Geist and F. Walther, eds. The behavior of
ungulates and its relation to management.
Vol. II, IUCN, Morges, Switzerland.
Cowan, I. McT., and W. McCrory.
1970. Variation In the mountain goat Oreamnos
americanus (Blainville). J. Mammal. 51(l):60-73.
66
DeBock, E.
1970. On the behavior of the mountain goat in
Kootenay National Park, British Columbia. M.S.
thesis, Univ. Alberta, Edmonton, Canada.
Denniston, R. H.
1956. Ecology, behavior and population dynamics
of the Wyoming or Rocky Mountain moose, Alces
alces shirasi. Zoologica 41:105-118.
Driver, C. H. , V. Stevens, and D. K. Pike.
1979. Terrestrial baseline surveys, non-native
mountain goats of the Olympic National Park.
Annual Report, 1978. Univ. Wash., Coll. For.
Resour. , Seattle.
Ewbank, R.
1967. Nursing and suckling behaviour amongst
Clun Forest ewes and lambs. Anim. Behav. ,
15:251-258.
Geist, V.
1965. On the rutting behavior of the mountain
goat. J. Mammal. 45:551-568.
Lent, P. C.
1974. Mother-infant relationships in ungulates,
p. 14-55. In_ V. Geist and F. Walther, eds. The
behaviour of ungulates and its relation to man-
agement, Vol. I, IUEN, Morges, Switzerland.
Li, A. K. C, M. J. Koroly, M. E. Schattenkerk,
R. A. Malt, and M. Young.
1980. Nerve growth factor: acceleration of the
rate of wound healing in mice. Proc. Natl. Acad.
Sci., USA. 77(7): 4379-4381.
Olmstead, I.
1976. Alpine and subalpine vegetation under the
influence of non-native mountain goats, Olympic
National Park. p. 1143-1148. I_n Proceedings of
the first conference on scientific research in
the national parks, Vol. II, U.S. Department of
the Interior.
Petocz, R. G.
1973. The effect of snow cover on the social
behavior of bighorn rams and mountain goats.
Can. J. Zool. 51:987-993.
Geist, V.
1971a. A behavioral approach to the management
of wild ungulates. p. 413-424. In E. Duffey
and A. S. Watts, eds. The scientific management
of animal and plant communities for conservation.
Blackwell Publ. , Oxford.
Geist, V.
1971b. Mountain sheep, a study in behavior and
evolution. Univ. Chicago Press. 383 p.
Hanson, W. 0.
1950. The mountain goat in South Dakota. Ph. D.
diss. Univ. Mich. , Ann Arbor.
Hanwell, A., and M. Peaker.
1977. Physiological effects of lactation on the
mother. Symp. Zool. Soc. Lond. 41:297-312.
Holroyd, J. C.
1967. Observations of Rocky Mountain goats on
Mt. Wardle, Kootenay National Park, British
Columbia. Can. Field Nat. 81(l):l-22.
Hutchins, M.
1980. Offspring retention in ungulates; a repro-
ductive strategy? Unpublished paper presented to
the 1980 Annual Meeting of the Animal Behavior
Society, Fort Collins, Colo.
Rideout, C. B.
1974. A radio-telemetry study of the ecology and
behavior of the mountain goat in western Montana.
Ph. D. thesis, Univ. Kans. , Lawrence.
Rideout, C. B. , and R. S. Hoffman.
1975. Oreamnos americanus. Maram.
63:1-6.
Species.
Sinclair, A. R. E.
1977. The African buffalo, a study of resource
limitation of populations. Univ. Chicago Press.
Tabor, R.
1975. Guide to the geology of the Olympic
National Park. Univ. Wash. Press, Seattle.
Trivers, R. L.
1974. Parent-offspring conflict. Amer. Zool.
14:249-264.
Tyler, S. J.
1972. The behavior and social organizations of
the New Forest Ponies. Anim. Behav. Monogr.
5(2):85-196.
Walker, E. P.
1975. Mammals of the world, Vol. II. John
Hopkins Univ. Press. London, England.
Kenny, J. T. , M. L. Stoloff, J. P. Bruno, and
E. M. Blass.
1979. Ontogeny of preference for nutritive vs.
non-nutritive suckling in albino rats. J. Comp.
Physiol. Psychol. 93(4) : 752-759.
Wilson, D. E., and S. M. Hirst.
1977. Ecology and factors limiting roan and
sable antelope populations in South Africa.
Wildl. Monogr. 54:1-111.
67
APPENDIX I
Behavior
Code
Description
Mother-offspring behaviors:
Nursing
A nursing attempt was defined as an approach and attempt to
suckle from the mother by her offspring. Attempts were usually
preceded by a short, rapid rush by the offspring toward the
udder. When this event was followed by contact with the udder,
it was classified as an attempt. An attempt was considered
"successful" only if the offspring remained in contact with the
udder for 23 seconds.
Approach modes
The kid approaches the udder from the mother's side and on the
same side on which it had been positioned just before the event
occurred.
Termination modes
Allogrooming
FR The offspring approaches the udder from between the mother's
front legs.
R The offspring approaches the udder from between the mother's
rear legs.
RA The offspring approaches the udder from the mother's side, but
before doing so, runs directly in front of her and makes udder
contact from the opposite side from which it began the approach.
SO The mother terminates a nursing attempt by lifting her rear leg
up over the offspring and walking forward.
WA The mother terminates a nursing attempt by walking forward,
without lifting the rear leg.
RLS The mother terminates a nursing attempt with a quick kicking
motion of the rear leg.
JA The mother terminates a nursing attempt by leaping away from the
offspring.
0 Any other termination. Descriptions were recorded in the form
of qualitative notes.
GR The licking of some portion of the body of another individual.
Agonistic Behavior:
Horn present
HP
Horn swipe, without contact HSW
Horn swipe, with contact HSC
The subject pulls in its chin and lowers the head, tilting the
horn tips slightly forward while pointing them in the direction
of another individual (Geist 1965, DeBock 1970).
The subject tucks in the chin, lowering the horn tips as in the
HP, then proceeds to sweep them upwards in a half arc in the di-
rection of another individual without making any physical contact
(Geist 1965, DeBock 1970).
Same as the preceding category, except that the animal performing
the behavior strikes some portion of another individual's body
with the horns.
Butt
Avoidance
BT
AV
The subject makes physical contact with another individual by
striking it with the anterior portion of the head and horns,
without, however, making the upward sweeping motion as is char-
acteristic of the horn swipe.
Any avoidance behavior associated with the approach or threat
behavior of another Individual. This may include crouching
and/or leaning in the opposite direction of a threatening indi-
vidual or, more commonly, actual movement away from the antago-
nist upon being approached or threatened.
68
Home Range and Habitat Use by Non-Migratory Roosevelt Elk,
Olympic National Park
Kurt J. Jenkins and Edward E. Starkey
ABSTRACT
Radiotelemetry was used to describe the home ranges, habitat use and social behavior of Roosevelt elk
(Cervus elaphus roosevelti) in the Hoh Valley of Olympic National Park. Cow elk were found to be
nonmigratory and home ranges were influenced primarily by topographic features. Flood plain areas were
selected during most seasons, but use of south-facing valley slopes increased during winter. Cow elk
formed relatively stable associations of adult females and their offspring. Olympic National Park
provided a unique opportunity to study native ungulates in undisturbed habitat, and additional research is
recommended.
Kurt J. Jenkins and Edward E. Starkey, Cooperative
Park Studies Unit, School of Forestry, Oregon
State University, Corvallis, Oregon 97331.
69
Figure 1. — Geographical location of Olympic
National Park and the Hoh valley study area.
INTRODUCTION
This research is part of a long-term investigation
of the behavior and ecology of Roosevelt elk
(Cervus elaphus roosevelti Merriam) in Olympic
National Park. The objective of this first phase
of the research was to describe the distribution
and movement of cow elk in a representative
watershed near the boundary of the Park. A pri-
mary purpose was to determine the extent to which
elk within Olympic National Park made use of
non-Park land. That information was needed to
assess effects of hunting and forest management
outside the Park on Park elk. A second goal was
to provide baseline information so that changes in
the distribution and behavior of elk within the
Park can be identified.
THE STUDY AREA
Olympic National Park occupies the central
mountainous portion of the Olympic Peninsula in
northwest Washington (fig. 1). The study area is
located in the valley of the main fork of the Hoh
River and extends from Canyon Creek, 3 km west of
the National Park boundary, upriver approximately
15 km. A single paved road extends the length of
the study area along the north side of the Hoh
River. The area west of the National Park bound-
ary is managed for timber production by the
Washington Department of Natural Resources.
Climate in the Hoh Valley is maritime, with mild,
wet winters and cool, dry summers. Average annual
precipitation is 345 cm. Areas below 600 m receive
mostly rain during winter, although approximately
25 cm of snow falls in valley bottoms each winter.
The Hoh valley has the broad U-shaped configura-
tion characteristic of glaciated watersheds.
Elevations range from 150 m on the valley floor to
910 m on adjacent ridgetops. The valley floor is
1.0 - 2.0 km wide and consists of gravel bars, at
least four river terraces, and various glacial
deposits. The vegetation has been described by
Fonda (1974) and represents a sequence of primary
succession from bare gravel adjacent to the river,
to mature Sitka spruce-western hemlock forests on
older terraces (fig. 2). Gravel bars, periodi-
cally flooded by the Hoh River, support pioneer
communities of young red alder (Alnus rubra) and
willow (Salix spp. ) . The youngest river terrace
is an alluvial deposit 80-100-years-old and
supports a mature red alder community. The next
oldest terraces, 400 to 700 years of age, also are
alluvial deposits and support serai Sitka spruce-
black cottonwood (Populus trichocarpa) and mature
Sitka spruce-western hemlock communities, respect-
ively. The oldest most extensive terrace is a
pleistocene glacial deposit, and it too supports
the climax Sitka spruce-western hemlock forest.
The valley bench, also a pleistocene deposit,
supports an additional climax forest characterized
by a dense canopy of western hemlock.
The spruce-hemlock forest of the Hoh valley,
commonly referred to as rain forest (Kirk 1966) or
moist coniferous temperate forest (Fonda 1974), is
a variation of the Picea sitchensis vegetation
zone, which occurs in coastal areas of Oregon and
Washington. Massive Sitka spruce and western
hemlock are common, and variable canopy coverage
results in a complex mosaic of understory vegeta-
tion. Small forest clearings are numerous and are
dominated by salmonberry (Rubus spectabilis) , vine
maple, and grasses. The shrub layers of denser
forest stands are dominated by huckleberry
(Vaccinium parvif olium, _V. alaskense) , vine maple,
and sword fern (Polysticum munitum). Overall,
heterogeneity is an important characteristic of
the spruce-hemlock climax forest.
Two additional plant communities on the valley
floor appear to be edaphically controlled. Bigleaf
maple communities occur on shallow, rocky soils of
alluvial fans formed by tributaries of the Hoh
River or of colluvial deposits at the base of the
valley wall. Vine maple communities occur on
alluvial outwashes as well as in areas that are
seasonally flooded by winter rains.
METHODS AND MATERIALS
Nine adult cow elk were immobilized and
radio-collared. They were immobilized by
injecting liquid or powdered succinylcholine
chloride into the hips of the animals, using a
powder charged Capchur™ rifle. Transmission
frequencies of radio-collars were approximately
164 MHz. The collars were distributed evenly
among three cow groups within the Park.
70
Figure 2. — Habitat units
within the Hoh valley
study area.
Collared elk were located during March 1978, from
1 June to 15 September 1978, and from 1 January to
20 March 1979, by triangulating from the ground
using an AVM LA-12 receiver and hand held yagi
antenna and by direct observation. The location
of a collared elk was established by determining
the direction of the strongest radio-signal from
three receiving stations and by plotting azimuths
on an orthographically corrected aerial photograph
(scale - 1:24,000). To minimize error, triangu-
lation stations were located on open gravel bars
where reflection of radio-signals was minimal. If
three bearings intersected, a circle was inscribed
within the resulting triangle. The center of the
circle was used as the estimated elk location and
assigned coordinate values. If bearings did not
intersect or if the accuracy of the location was
doubtful, telemetry equipment was used to locate
the elk visually or aurally. Each elk was located
one to three times daily usually at different
times each day.
Home range was defined as "the area over which an
animal normally travels in pursuit of its routine
activities" (Jewell 1966). Therefore, infrequent
journeys by elk beyond their normal range were
excluded from analyses. An elliptical home range
model (Koepple et al. 1975) was used to delineate
annual, home ranges of collared elk. Home ranges
lacked definite boundaries by that method and were
expressed as elliptical areas that included 95
percent of an animal's activity.
Daily movements were estimated each season by
measuring the line distance between an elk's
location one morning and its first location the
following morning. Undoubtedly, the actual
distance moved each day was greater than the
distance calculated. However, the index was
useful in examining seasonal changes in daily
movements.
Habitat was examined by comparing availability
within the composite home range of all radio-
collared elk to utilization (Neu et al. 1974).
Availability was defined as the percent of area
within composite home ranges covered by each
habitat unit and utilization as the percent of
radio locations within each habitat unit. Elk
selected a zone or habitat unit if utilization was
significantly greater (p<0.05) than availability
or avoided it if utilization was significantly
less than availability (p<0.05). Use of habitat
units was described during summer (10 June to 31
August 1978), winter (1 January to 28 February
1979) and early spring (1 March to 20 March 1978,
and 1979). Early spring was delimited because
there was a noticeable shift in habitat use at the
beginning of March.
Interactions between elk groups were described by
examining spatial overlap of home ranges and
association of collared elk within each group.
Association referred to the percent of time a
collared elk was located in the presence of each
other collared elk. Coefficients of association
were calculated according to the procedure
described by Cole (1949) and applied to elk by
Knight (1970) and Schoen (1977). Values range
linearly from 0.0 to 1.0, indicating no association
to perfect association, respectively. Inferences
concerning stability of elk groups were drawn by
comparing association of elk between groups.
RESULTS AND DISCUSSION
Approximately 180 elk inhabited the study area.
These were distributed among three groups of about
60 indivuduals each.
71
SCALE
2KM
I
N
1
•••••••••.....•
c*>
RANG!
.STATIC
Ni
>~
^
""o.
7VV
J VALLEY FLOOR
......... RIDGE
PARK BOUNDARY
I
$OUTHyrffi
"••..
"••••.../
•••"•
Mtllim
Figure 3. — Home ranges of cow elk in three home
range groups in the Hoh valley.
Home Range and Movements
Radio-collared elk were non-migratory. None
dispersed from the lowlands to summer range at
higher elevations; nor was there an observable
seasonal movement of elk across the National Park
boundary. No collared elk observed in Olympic
National Park was ever located more than 2.0 km
outside of the Park. Additionally, no elk from
more than 5.0 km within the Park was ever observed
on State-owned land adjacent to the Park.
Skinner (1933) nd Schwartz (1939) suggested that a
migratory portion of the population remained in
the upper Hoh River watershed unless deep snow
caused them to move down river. Under such
conditions, migratory elk may cross the National
Park boundary. It is unlikely, however, that
cutover areas outside the Park would be available
to elk during severe winters because of snow
accumulation. Elk would probably occupy densely
timbered hillsides within the Park, as observed by
Newman (1956). Therefore, it seems doubtful that
either resident or migratory elk in the Hoh valley
make use of non-Park land during winter. Informa-
tion on migratory elk in the Hoh valley still is
lacking.
Seasonal movement of elk up and down valley was
minimal. Two elk, however, traveled independently
to the Bogachiel valley to the north on 14 August
1978 and 17 August 1978. Each traveled 4.8 km
from its last confirmed location in the Hoh valley
and was absent from its home range for 12 to 18
days before returning. Although the cause was
unknown, the movements may have been associated
with reproductive behavior since they occurrred
near the onset of the rut. Lieb (1973) reported
that cow elk in Prairie Creek Redwoods State Park
occasionally left their group and wandered prior
to the onset of the rut. Hormonal changes during
estrus probably contribute to the restlessness of
cow elk (Lieb 1973). Movement of cow elk between
watersheds indicated that intermingling of popula-
tions from adjacent drainages may occur; however,
interchanges of non-migratory elk between
watersheds probably is uncommon. Only two elk
moved extensively during this study, and both
returned to their normal range.
A total of 2,565 locations of the radio-equipped
elk was obtained. Annual home range areas, during
the entire study period, average 1 112 ha. Orien-
tation and size of home ranges within the Park
were influenced by the valley floor (fig. 3). The
major axis of each ellipse was alined closely to
the flood plain of the valley. Additionally, width
of home range was related significantly to the
breadth of the valley floor measured through the
geometric center of the home range ( r^ = 0.57,
p<0.05). Home ranges were broadest near Twin Creek
where the valley floor was most extensive. In
southwest Oregon, home ranges of Roosevelt elk
also were influenced by topographic features.
Home range diameters averaged 1 . 3 mi in steep
72
Table 1 — Percent availability and use of vegetation types in the Hoh valley,
Olympic National Park.
Utilization
(percent of radio-relocations)
Availability
Location
(percent of
Vegetation
composite
type
home range)
Summer
Winter
Ea
rly spring
Gravel/willow
0.06
0.04(o)i
0.02(-)
0.01(-)
Red alder
.05
•09(+)
.10(f)
•25(+)
Spruce-cottonwood .02
.02(o)
.03(o)
•13(+)
Spruce-hemlock
.35
-53(+)
• 44(+)
.47(+)
Bigleaf maple
.04
•13(+)
•10(+)
.05(o)
Valley floor
Valley wall2
South
exposure Western hemlock.
North
exposure Western hemlock
Total
No. of radio-locations(n)
.23
•io(-)
.20(o)
•07(-)
.25
.09(-)
• IK")
•02(-)
1.00
1.00
1.00
1.00
683
759
541
■'-Symbols indicate significant habitat selection (+) , avoidance (-), and
neutrality (o) based on family confidence intervals using Bonferroni Z-statistics
(Neu et al. , 1974).
9
Represents only the lower one-third of the valley wall (150-225 m of elevation)
because upper two-thirds were not within composite home ranges.
canyons; whereas in valley floor plains, they
averaged 3.0 mi (Harper 1971).
Daily movements of cow elk were greater during
summer than winter (p<0.05). Elk traveled an
average minimum distance of 843 m between succes-
sive mornings in summer and 688 m in winter. This
reduction of activity in winter may conserve
energy at a time when weather is typically severe
and forage does not satisfy maintenance require-
ments. The metabolic rates of black-tailed deer
(Odocoileus hemionus columbianus) , mule deer (0. h.
hemionus) , and white-tailed deer (0_. virginianus)
fluctuate annually with fasting heat production
highest during summer and lowest during winter
(Regelin 1979, Silver et al. 1969, Thompson et al.
1973). It seems likely that metabolic rates of
elk are reduced similarly during winter, resulting
in decreased movement and home range size.
Daily movements of cow elk with calves were least
in June, traveling an average minimum distance of
541 m/day in June; whereas those without calves,
traveled an average of 1 040 m. Movement of cows
with calves was restricted in June because
frequent nursing required that cows stay close to
the less mobile calves.
Darling (1937) emphasized the importance of
tradition as a determinant of movement and home
range use of red deer (£. e. elaphus). Similarly,
a high level of traditionality appeared to be
associated with home range use by elk in the Hoh
valley. Several heavily used foraging areas and
travel corridors existed within home ranges and
were used in a sequential, and often predictable
pattern. Generally, elk remained in a favored
foraging area for 1 to 3 days. Although elk often
moved through their home ranges in a predictable
manner, an overall circuit of travel did not exist.
Patterns of Habitat Use
Elk selected valley floor zones during all seasons
and generally avoided valley walls (table 1). The
valley floor was probably the most important habi-
tat because the assemblage of plant communities
provided the best dispersion of forage and cover.
Clearings on the forest floor were used heavily by
feeding elk, generally sustained the highest
browsing intensity, and probably provided more
abundant and diverse forage than valley walls.
Although the valley floor was important in all
seasons, elk used lower parts of the valley wall
under certain conditions. Use of the lower north
valley wall (south-facing) increased during winter
1978 (table 1) and was frequented by elk during
clear, cold periods and when snow cover existed.
South facing slopes intercept more solar radia-
tion, which results in warmer temperatures and
lower snow depths than on the valley floor. The
relatively dense canopy on valley walls may inter-
cept more snow and contribute to reduced snow
depths (Jones 1974). At night, significantly more
73
1
1
6
9
6 10
Z 11
* 88
u 4
7
12
3
8
100
094
too
0.76
071
1.00
000
0.00
0.00
100
Entire Study
001
0.01
000
0.75
100
0.02
002
0.01
0.88
088
1.00
0.01
001
0.01
1.00
001
0.01
0.49
1.00
0.01
0.01
0.01
088
1.00
0.01
0.01
0.00
1.00
0.05
0.01
0.00
0.t3
1.00
1
6
9 10 11 88 4 7 12 3 8
ELK NO
1
6
9
6 10
z
11
5 4
ta* 7
3
8
1.00
0.94
1.00
0.85
092
1.00
June Onlv
0.00
0.00
0.00
1.00
0.00
0.00
000
0.26
1.00
0.00
0.00
1.00
0.00
0.00
on
1.00
000
0.00
1.00
0.00
0.06
000
100
1 6 9 10 11 4 7 3 8
ELK NO
Figure 4. — Coefficients of association for pairs
of radio-collared elk in the Hoh valley. Shaded
cells represent pairs of elk. within a home range
group and non-shaded cells represent pairs from
adjacent groups.
downward infrared energy is radiated from a rela-
tively dense canopy than from either less dense
canopies or the clear sky (Moen 1973). Temperature
inversions probably are common, and together with
the above factors, may create more moderate thermal
conditions on the valley wall than on the valley
floor during winter.
Elk on the valley floor selected spruce-hemlock
and red alder habitat units during all three
seasons. Gravel bars were generally avoided;
their use was greatest during summer, especially
during hot midday periods in July. Breezes may
have provided relief from insects and summer
heat. Big leaf maple habitats were selected
during summer and winter, and used in proportion
to availability in spring. Spruce-cottonwood
habitats were used in proportion to availability
except in the spring, when they were selected.
The shift in utilization to red alder and spruce-
cottonwood habitats during early spring probably
was a response to abundant new growth of grass.
Grass is an important springtime dietary component
for elk (Schwartz and Mitchell 1945). Frequently,
elk traveled from these communities to adjacent
spruce-hemlock stands to bed during the midday
period.
Social Group Interactions
The individual home ranges of three collared elk
in each of the three groups coincided closely
(fig. 3). Although home ranges of elk from
adjacent groups overlapped, no permanent inter-
change of collared elk occurred between groups.
Elk from adjacent groups were observed together in
only eight cases. Association lasted for 3 days
or less (x = 1.4 days), and the original groups of
collared elk were preserved after temporary
associations. Coefficients of association for
pairs of elk within a group (x = 0.71, range =
0.13 to 0.94) were greater than those between
groups (x = 0.01, range = 0.00 to 0.05) (fig. 4).
The lack of perfect association between elk in a
group indicted that subgroups were periodically
absent from the main group. Duration of those
periods average 5.4 days (n = 42) but was highly
variable (S.D. = 6.2 days). The coefficients of
association between elk No. 7 and No. 8, which
were known to have calves, were lowest in June
(fig. 4). This may indicate subgrouping was more
common during calving.
Previous studies suggested that stability of elk
groups was variable. Darling (1937) believed that
groups of red deer consisted of a dominant female,
her mature daughters, and their offspring. The
matriarchal concept has been accepted widely
(Altmann 1952, McCullough 1969, Franklin et al.
1975). It implied that home range was passed on
through generations, a trait known as home range
conservation (Murie 1951), and that elk groups
were stable, comprised of a constant membership.
In contrast, Knight (1970) found that the mean
coefficient of association between female Rocky
Mountain elk (C. e. nelsoni) never exceeded 0.47,
and felt that elk groups should be considered
aggregations rather than social groups. Schoen
(1977) reported a mean coefficient of association
of 0. 20 for 39 female Rocky Mountain elk in
western Washington, which also indicated low group
stability. Marcum (1975), Mackie (1970), and
Shoesmith (1979) reported that Rocky Mountain elk
groups changed composition frequently; however,
the effect of seasonal migration and hunting on
the constancy of those groups is unknown.
Harper (1971) found that Roosevelt elk herds on
managed forest lands in southwest Oregon contin-
uously changed composition and that marked members
of adjacent groups interchanged freely. In
contrast, non-migratory and unhunted elk in the
unmanaged Prairie Creek Redwoods State Park formed
more stable associations of adult females and
their immature offspring (Franklin et al. 1975).
In that population, small fluctuations in group
size occurred as subgroups entered and left the
herd; however, absent individuals always
returned. Franklin and Lieb (1979) hypothesized
that variations in group stability of non-migratory
elk in Prairie Creek Redwoods State Park and in
southwest Oregon may be caused by differences in
habitat. They suggestd that the continual altera-
tion of vegetation due to logging in southwest
Oregon created changing habitat conditions which
affected the development and maintenance of social
organization. They concluded that relatively
74
stable social organizations may be expected in elk
populations that inhabit stable, unraanaged environ-
ments which permit long-term bonding among
individuals.
SUMMARY AND CONCLUSIONS
Home ranges of elk were influenced primarily by
topographic features of the Hon valley. Major
axes of each home range were alined to the flood
plain of the valley, and widths of home ranges
were related to the breadth of the valley floor.
Additionally, radio-locations of the collared elk
were concentrated on the valley flood plain. It
seems likely that the distribution of elk in the
unmanaged forest setting is governed largely by
the location of suitable foraging areas. Numerous
forest clearings on the flood plain provide
environmental heterogeneity and abundant forage
for elk. Thus, the extent of valley floor may be
an important determinant of elk density in
unmanaged fluvial forests on the Olympic Peninsula.
Current logging practices are sometimes suggested
to be beneficial to elk because the removal of the
overstory trees creates new foraging areas for
elk. We found, however, that use of densely
timbered, southfacing walls increased in winter,
especially during cold periods with snow cover.
Closed canopy forests on south-facing slopes may
be extremely important habitat during periodically
severe winters, both because they provide a more
moderate thermal environment and because lower
snow depths may result in greater availability of
forage (Jones 1974). Although the relative value
of mature forests and clearings as foraging areas
on elk winter range remains poorly understood,
managers should be aware of the importance of
mature forest under certain winter conditions.
Seasonal differences in phenology and thermal
regimes appear to have influenced distribution of
elk in the Hoh valley. Additionally, dietary
preferences probably influenced selection of
habitat units on the valley floor.
During summer and winter, relatively more use
occurred on the older river terraces and valley
walls. During early spring, use shifted to
riparian alder-flat and spruce-cottonwood
habitats. This appeared to be in response to new
growth of grasses. Late winter is a crucial
period for cow elk nutritionally. Protein and
energy reserves are low after winter, and demands
of rapid fetal growth are high (Moen 1973). Thus
red alder and spruce-cottonwood communities should
be considered critical elk range in the fluvial
rain forest valleys.
Cow elk formed relatively stable associations of
adult females and their offspring and conformed to
the model of group behavior of Franklin and Lieb
(1979). There was no permanent interchange of
collared elk between groups, and elk within groups
were more highly associated than reported else-
where. These findings support the hypothesis that
more stable elk groups may form where habitat is
constant and not disturbed by logging. Additional
comparative research on the behavior of elk groups
in managed and unmanaged forest settings is needed
to further test the hypothesis.
Apparently, non-migratory cow elk in Olympic
National Park are influenced little by forest and
wildlife management practices that occur outside
the Park. Park elk within 5.0 km of the Park
boundary made use of non-Park land and thus, may
be hunted and influenced by habitat changes which
occur there. Beyond a 5.0-km strip, elk were
isolated from events outside the Park. Therefore,
management practices outside the Park do not
affect the majority of cow elk within the Park and
may not be used to manage them. Information on
migratory and male segments of the Hoh valley
population is needed before the overall influence
of external factors on elk in Olympic National
Park can be described completely.
Results of this study also indicate that because
the majority of elk in the Hoh valley are not
influenced by events outside the Park, the behav-
ior and ecology of elk within the Park are the
most accurate reflection of the primeval condition
of elk on the Olympic Peninsula. These elk provide
a unique opportunity to research natural regula-
tion of herbivore populations and the influence of
elk on forest communities. Additionally, compara-
tive research on elk within and outside of Olympic
National Park is important to determine the
influence of forest management on elk movement,
productivity, and social organization.
LITERATURE CITED
Altmann, M.
1952. Social behavior of elk (Cervus canadensis
nelsoni) in the Jackson Hole area of Wyoming.
Behavior 4:116-143.
Cole, L. C.
1949. The measurement of interspecific
association. Ecology 30(4) : 411-424.
Darling, F. F.
1937. A herd of red deer.
Press, London. 215 p.
Oxford University
Fonda, R. W.
1974. Forest succession in relation to river
terrace development in Olympic National Park,
Washington. Ecology 55:927-942.
Franklin, W. L. , A. S. Mossman, and M. Dole.
1975. Social organization and home range of
Roosevelt elk. J. Mammal. 56; 102-118.
Franklin, W. L. , and J. W. Lieb.
1979. The social organization of a sedentary
population of North American elk: a model for
understanding other populations, p. 185-198. Ln
M. S. Boyce and L. D. Hayden-Wing, eds. North
American elk: ecology, behavior and management.
Univ. of Wyo. , Laramie. 294 p.
Harper, J. A.
1971. Ecology of Roosevelt elk. Oreg. State
Game Comm. , Portland, OR. P-R Proj. W-59-R.
44 p.
Jewell, P. A.
1966. The concept of home range in mammals.
Symp. Zool. Soc. Lond. 18:85-109.
75
Jones, G. W.
1974. Influence of forest development on
black-tat led deer range on Vancouver Island,
p. 139-148. In H. C. Black, ed. , Wildlife and
forest management in the Pacific Northwest.
Oregon State University, Corvallis. 231 p.
Kirk, R.
1966. The Olympic rain forest.
Press, Seattle. 86 p.
Knight, R. R.
1970. The Sun River elk herd.
23:1-65.
Univ. Wash.
Wildl. Monogr.
Koepple, J. W., N. A. Slade, and R. S. Hoffman.
1975. A bivariate home range model with
possible application to ethological data
analysis. J. Mammal. 56:81-90.
Lieb, J. W.
1973. Social behavior in Roosevelt elk cow
groups. M.S. thesis. Humboldt State Univ.,
Areata. 82 p.
Mackie, R. J.
1970. Range ecology and relations of mule deer,
elk, and cattle in the Missouri River Breaks,
Montana. Wildl. Monogr. 20:1-79.
Marcum, C. L.
1975. Summer-fall habitat selection and use by
a western Montana elk herd. Ph.D. thesis.
Univ. Mont. , Missoula. 294 p.
McCullough, D. R.
1969. The tule elk: its history, behavior and
ecology. Univ. Calif. Publ. Zool. 88:1-209.
Mo en, A. N.
1973. Wildlife ecology. W. H. Freeman and Co. ,
San Francisco, Calif. 458 p.
Murie, 0. J.
1951. The Elk of North America. The Stackpole
Co., Harrisburg, Pa., and Wildlife Management
Institute, Washington, D.C. 386 p.
Neu, G. W. , C. R. Byers, and J. M. Peek.
1974. A technique for analysis of utilization-
availability data. J. Wildl. Manage. 38(3):
541-545.
Newman, C. C.
1956, Biologist's annual report — Elk.
Unpublished manuscript. Olympic National Park,
Port Angeles, Washington.
Regelin, W. L.
1979. Nutritional interactions of black-tailed
deer with their habitat in southeast Alaska,
p. 60-68. In 0. C. Wallrao and J. W. Schoen, eds.
Sitka black-tailed deer: proceedings of a
conference in Juneau, Alaska. USDA Forest
Service, Series No. R10-48. Juneau, Alaska.
231 p.
Schoen, J. W.
1977. The ecological distribution and biology
of wapiti (Cervus elaphus nelsoni) in the Cedar
River watershed, Washington. Ph.D. thesis.
Univ. Wash., Seattle. 408 p.
Schwartz, J. E.
1939. The Olympic elk study. Unpublished
manuscript. U.S. Forest Service, Olympia, WA.
Schwartz, J. E., and G. E. Mitchell.
1945. The Roosevelt elk on the Olympic
Peninsula, Washington. J. Wildl. Manage.
9:295-319.
Shoesmith, M. W.
1979. Seasonal movements and social behavior of
elk on Mirror Plateau, Yellowstone National Park,
p. 166-176. InM. S. Boyce and L. D.
Hayden-Wing, eds., North American elk: ecology,
behavior, and management. Univ. of Wyo. ,
Laramie. 294 p.
Silver, H. , N. F. Colovos, J. B. Holter,
and H. H. Haynes.
1969. Fasting metabolism of white-tailed deer.
J. Wildl. Manage. 33:490-498.
Skinner, M. P.
1933. Report on elk conditions in Olympic
Mountains to Boone and Crockett Club.
Unpublished manuscript. Olympic National Park,
Port Angeles, Washington.
Thompson, C. G. , J. B. Holter, H. H. Haynes,
H. Silver, and W. E. Urban, Jr.
1973. Nutrition of white tailed deer I. Energy
requirements of fawns. J. Wildl. Manage.
37:301-311.
76
Pollutant Monitoring in the Olympic National Park Biosphere Reserve
Kenneth W. Brown and G. Bruce Wiersma
ABSTRACT
Scientific interest in global contamination has been instrumental in the establishment of 33 Biosphere
Reserve sites throught the United States. These sites, including many pristine areas that are and have
beer, protected from industrial development, serve as areas in which present and future environmental
pollution can be assessed.
The Olympic National Park Biosphere Reserve was selected by the National Park Service as the second U.S.
site, following Great Smoky Mountains National Park, for pollutant monitoring studies. These studies,
conducted by the U.S. Environmental Protection Agency, were designed to identify levels of trace elements
and organic contaminants in both the physical and biological media. Based on the Great Smoky Mountains
experience, 10 remote sites within the Park were selected for intensive sampling. Sampling areas were
located in the Hoh, Quinault, and Dosewallips River drainages; at Anderson and Grand Pass; and near the
northern most edge of Blue Glacier. Their proximity to vehicle-traveled roads varied from 5 to 12 miles.
An additional sampling area was located close to Ozette Lake near the Pacific Ocean.
The media sampled included air, water, soils, litter, and several different plant species. These samples
were processed and analyzed for selected organic and 26 different heavy metal contaminants such as lead,
cadmium, nickel, and zinc. Additional analyses included airborne particulate characterization, such as
size determinations and the identification of chemical constituents.
The data obtained identify then current baseline contaminant concentrations and will assist in resource
management and environmental quality programs.
Kenneth W. Brown and G. Bruce Wiersma, Exposure
Assessment Research Division, Environmental
Monitoring Systems Laboratory, U.S. Environ-
mental Protection Agency, Las Vegas, Nevada
89114.
77
INTRODUCTION
In August 1979, a study was initiated to identify
pollutant concentrations in specific biological and
physical media in Olympic Nation.il Park for the
development of a pollutant monitoring system.
This study, similar to that conducted in Great
Smoky Mountains National Park as described by
Wiersma et al. (1979), was a cooperative effort
between the U.S. Environmental Protection Agency's
Environmental Monitoring Systems Laboratory at
Las Vegas (EMSL-LV) and the National Park Service
(NPS).
Olympic National Park, a public reserve of approxi-
mately 360 000 ha (889,000 acres), is one of the 33
designated biosphre reserves in the United States.
The origination of the Biosphere Reserve System
and the subsequent selection of Olympic National
Park and other specific Biosphere Reserve sites
were initiated in part because of a concern for
United States and global contamination from human
industrial activities.
The concept and criteria used to establish this
reserve system have been described previously by
many researchers including Franklin (1977) and
Wiersma and Brown (1979). Basically, the Biosphere
Reserve sites are physically and biologically
undisturbed and protected natural background areas
where life processes occur with minimal human
interference. They are of value to concerned
environmentalists and scientists because they:
1. provide permanent and undisturbed areas where
long-term background or baseline studies can be
conducted on environmental and biological features;
2. aire natural sources of genetic pools of animal
and plant species;
3. provide areas for assessing, identifying, and
recording the physical and biological state of the
environment ;
4. provide endemic habitat to obtain data from
local environmental studies instrumental to the
formation of management plans and policies for the
Reserve;
5. provide areas for long-term biological
research; and
6. serve as sites for measuring and assessing the
concentration and impact of human made pollutants
on biological systems.
Use of Biosphere Reserves as pollutant monitoring
sites originated with the Man and Biosphere
program (MAB) at the 16th General Conference of
the United Nations Educational, Scientific and
Cultural Organization (UNESCO) and with a 1970 ad
hoc task force concerned with the Global Network
for Environmental Monitoring (GNEM). The recom-
mendations, criteria, and coordination of these
groups of concerned scientists have been described
previously (Wiersma and Brown 1979).
The experimental design, methods, and approach
used for collecting data to develop a pollutant
monitoring system for Olympic National Park
followed the basic principle of the monitoring
systems design described by Schuck and Morgan
(1975) and Morgan et al. (1979). The approach
used addresses both the multimedia concept
(sampling and analysis of pertinent biological and
physical media) and the systems concept (identifi-
cation of the interaction and/or kinetics between
media). The methods for assessing and addressing
the systems concept using kinetic modeling have
been described by O'Brien (1979) and Barry (1979).
Wiersma (1979) applied this method to the analysis
of a monitoring project conducted by the Environ-
mental Protection Agency, Environmental Monitoring
Systems Laboratory, Las Vegas, and National Park
Service in Great Smoky Mountains National Park.
This report describes and identifies the sampling
locations, media collected, and equipment used in
Olympic National Park. Objectives were to deter-
mine pollutant levels, assess variability between
collected media, determine biological accumulators
of selected contaminants, and evaluate sampling
equiment. Media sampled included air, litter,
water, vegetation, and soil.
SAMPLING METHODS
To sample adequately this biologically diverse
National Park, seven multimedia and eight single
media locations were selected. All the sites were
chosen to obtain data representative of a partic-
ular set of relatable biological and/or physical
parameters. For example, multimedia sampling
locations were selected in areas impacted and
influenced by ocean environments. In addition,
high and low altitude sampling locations, repre-
sentative of indigenous forest communities and
influenced by both regional and local meteor-
ological conditions, were selected. Single media
sampling locations were established to provide
pollutant data on the major incoming and outgoing
routes of exposure. These included high altitude
air sampling sites and water sampling locations in
the Park's major drainage systems.
The major route of exposure for this Park is
considered to be the air pathway. As such,
emphasis was placed upon the methods and means to
collect and measure airborne contaminants. This
is especially important in Olympic National Park,
in that, even though the predominant air masses
originate to the north and northwest, local topog-
raphy influences the air flow to a great extent.
According to B. Moorhead (Research Biologist,
Olympic National Park, Port Angeles, Washington,
personal communication, 1979), the air flow is
affected by the conical shape of the local ter-
rain. This phenomenon can create problems in
assessing contaminant impact from large incoming
air masses versus that from possible transport and
redeposition by local meteorological conditions.
The techniques and procedures for
Biosphere Reserves have been descr
by Brown et al. (1979). Equipment
Olympic National Park study includ
air sampling devices. The pump, a
air pump, is capable of a maximum
cnW/min. The power sources are Ga
acid batteries connected in parall
capable of operating the pump for
days.
sampling air in
ibed previously
used in the
ed a pump and
Dupont P400A
flow of 4000
tes^ sealed
el that are
10 continuous
^Registered trademark.
78
The air sampling devices used Co collect samples
at each air sampling site consisted of combinations
of both plastic Millipore aerosol monitors and
stainless steel holders. Both types were loaded
with 47-mm-diaraeter Millipore filter membranes
with a 0.45-um pore size.
Collection of four air samples per site was due in
part to the three different analytical procedures
used to identify trace-element contamination:
X-ray fluorescence, atomic absorption spectrometry,
and scanning electron microscopy (SEM). The fourth
sample was archived in case of loss or for further
analytical requirements. Prior to use, all stain-
less steel holders and filter membranes were
cleaned and packed in a clean laboratory facility.
The clean laboratory techniques (C. Davidson,
Carnegie-Mellon Univ., Pittsburg, Pa., personal
communication, 1979) were necessitated because of
the low flow rates which varied between 900 and
1 000 cm-Vmin per sample and the expected low
elemental concentrations at each of these sampling
sites (fig. 1).
Contamination of air samples was minimized after
collection by wrapping each filter in aluminum
foil and placing each in a plastic bag. The
samples were then transported to the appropriate
laboratory for analysis.
Dry fall deposition of airborne particulates was
collected and examined on Teflon plates at 10 of
the sampling locations. The simultaneous measure-
ments of airborne concentrations and Teflon plate
concentrations allow calculations of deposition
velocities for a number of contaminants. Clean
laboratory procedures also were used for the
deposition plate components.
Water samples were collected by two different
methods. At each sampling site, a 250-ml grab
sample was collected from a nearby stream. The
sample was placed in a precleaned (acid-washed)
polyethylene bottle and fixed with 1 ml of ultrex
nitric acid. At four sites, a gravity flow resin
column was used. This column, packed with
selected resin filters for entrapping selected
organic compounds, was placed in the river or
stream to be sampled. Natural water flow forced
water through the column. The resins selected for
this water sampling device are described by
Wiersma et al. (1979).
Vegetation, soil, and litter were collected at
seven locations (fig. 1). The design and number
of samples collected were based upon previously
defined factors and limitations as described by
Wiersma et al. (1979) and Wiersma and Brown (1979).
At each collection location, five different plant
species were taken from 10 different subsites.
Thus, a total of 50 plant samples, representing 10
replications of 5 different species, was collected
for analysis. In addition, one soil sample
(0-5 cm in depth) and one litter sample was
collected from each of the subsites, making a
total of 10 samples each per sampling location.
The 50 plant, 10 soil, and 10 litter samples were
placed in clean polyethylene bags and sealed. All
were analyzed for trace elements. One soil and
one litter sample collected from one of the
subsites were placed in clean 1-liter Teflon
bottles for organic analysis.
• AIR (HIGH ALTITUDE)
▼ WATER
OvEGETATION-SOIL-LITTER (MULTIMEDIA)
^SNOW
Figure 1. — Sampling locations in the Olympic
National Park.
Rain gauges were installed at the seven plant- and
soil-collection locations to collect and measure
any precipitation. No precipitation fell at these
sites during the collection period.
SAMPLING LOCATIONS
Single Media Air Sampling Sites
Three different sites were selected for high alti-
tude air sampling. As shown in figure 1, the
sites are fairly evenly distributed in the higher
elevations of the Park. The first site was located
on the ridge above Moose Lake, with access via the
ridge line or by the Grand Pass Trail from
Hurricane Ridge. The second site was located
about 900 m south of the Glacier Meadows Ranger
Station near the eastern edge of the Blue Glacier
moraine. The third site was located about 1 800 m
west of Anderson Pass and 400 m south of the
Anderson Pass Trail.
The following criteria were used for site
selection: the site should be
1. free of local contamination and at least 5 km
from the nearest road used by automobiles and
other vehicles;
2. located at as high an elevation as possible,
and
3. located in a clearing in which the diameter is
at least five times the height of the surrounding
forest.
The sampling heads at these three sites were
supported by aluminum stands. The heads were
1.5 m from the ground. The pumps, battery, and
flow meters were placed on the ground during
operation.
79
Single Media Water Sampling Sites
Plant species collected included
Four different major drainage systems were chosen
tor the resin filter collectors. Grab samples, as
previously described for elemental analysis, also
were collected. The drainage systems selected for
sampling, as shown in figure 1, included the Elwha,
Hoh, Quinault, and Dosewallips. These four drain-
age systems drain slightly more than 50 percent of
the Park's watersheds. Length of sampling varied
from 5 to 8 days.
In addition to water samples, surface snow samples
were collected from the Blue Glacier for trace
element analysis. Location of this collection
site, as shown in figure 1, was the eastern edge
of the glacier near its midpoint.
Multimedia Sampling Sites
At these sites (fig. 1), air, water, vegetation,
soil, and litter were collected. Air monitoring
equipment and techniques used for sampling air in
these locations were identical to those used for
the single media air sampling sites. The only
difference was that the filter holders were placed
under a canopy and supported by young saplings.
The number of samples and the techniques used for
sampling vegetation, soils, and litter were
described previously. In addition, deposition
plates were placed near the air sampling equipment
and grab samples of water were collected from each
location.
Low Altitude Hoh
On the north side of the Hoh River a
1 800 m west of the Olympus shelter,
dominated by Sitka spruce, Picea sit
western hemlock, Tsuga heterophylla.
by Franklin and Dyrness (1973), the
characteristics of this forest are (
of Acer macrophyllum and A. circinat
spicuous coverage of epiphytic plant
the most abundant being Selaginella
club moss; (3) abundant nurse logs;
tively great densities of Roosevelt
both seasonal and resident).
pproximately
the site is
chensis , and
As described
distinctive
1) an abundance
um; (2) con-
s, with one of
oregana, a
and (4) rela-
elk (elk are
The major herbaceous species are Oxalis oregana,
Polystichum muni turn, Tlarella unlf oliata, Carex
deweyana , Trisetum cernuum, Malanthemum dilatatum,
Rubus pedatus, Montia sibirica, Athyrium
f llix-f emina , and Gymnocarpium dryopteris;
Polystichum and Oxalis are clearly the most
important. A heavy moss layer is typical includ-
ing Eurhynchium oreganum, Hypnum circinale ,
Rhy tidiadelphus loreus, Leucolepis menziesii ,
Plagiomnium insigne , and Hylocomium splendens as
more common species. The heavy epiphyte coverage
includes the cryptogams Isothecium stolonlf erum,
Porella navicularis, Rhy tidiadelphus loreus,
Radula bolanderi , Frullania nisquallensi s ,
Scapanla bolanderi , and Ptilidium californicum,
and the vascular plants Polypodium vulgare and
Selaginella oregana. This vegetation type
occurring on the second terrace stage is further
characterized by areas of shallow stony soil.
1.
2.
Acer circinatum
Tiarella trifoliata
vine
foam
wood
swore
moss
maple
flower
3.
4.
Oxalis oregana
Polystichum munitum
sorrel
fern
5.
Rhy tidiadelphus
loreus
High Altitude Hoh
Located about 2 000 m north of the Glacier meadows
shelter just west of the Hoh trail, this sampling
area is dominated by mountain hemlock, Tsuga mer-
tensiana with areas of silver fir Abies amabilis.
This vegetative type usually dominates the higher
forested zones along the western slopes of the
Olympic National Park.
Chamaecyparis nootkatensis is a major associate,
with Pseudotsuga menziesii , Abies lasiocarpa, and
Pinus monticola as minor associates. Of the wide
variety of understory species, many belong to the
Ericaceae, Rosaceae, and Compositae families.
Soils are podzolic; however, the degree of podzol-
ization varies depending on location.
oval-leaf huckleberry
thin-leaved huckleberry
Alaska huckleberry
blackberry
Species collected included:
1. Vaccinium ovalifolium
2. Vaccinium membranaceum
3. Vaccinium alaskaense
4. Rubus pedatus
5. Streptopus streptopoides twisted-stalk
Ocean Site (Ozette)
This sampling site was located about 5 000 m
southwest of the Ozette trailhead to Sand Point.
Described by Franklin and Dyrness (1973), this
vegetation type is a western hemlock-Si tka spruce/
Gaultheria shallon, salal, association, commonly
found along the coastline and including a dense
shrub layer of salal. This species commonly con-
tributes over 40 percent of the species compositon.
Other common species include Oxalis oregana and
Thuja plicata.
Species collected included:
1. Moss species (genus and species unidentified)
2. Fern species (genus and species unidentified)
3. Lyslchitum-americanum skunk cabbage
4. Vaccinium ovatum evergreen huckleberry
5. Gaultheria shallon
salal
Low Altitude Quinault
About 2 000 m northeast of the Enchanted Valley
trailhead in the east fork of the Quinault River,
this sampling area was a forest type dominated by
Pseudotsuga menziesii and western hemlock, a type
that has a rich abundance of understory species
including grasses, ferns and a wide variety of
mosses. Located on a second-stage terrace, the
soils were well drained and well developed.
80
Table 1 — Analytical methods utilized for the analysis of the Olympic National Park
samples
Sample type
Analytical method
Analysis
Water
Water
Air
Vegetation
(includes
litter)
Soils
Gas chroraatograph-mass spectrometry
(GC-MS)
Inductively coupled plasma emission
(ICPES)
Spark source mass spectrometry
(SSMS)
X-ray fluorescence
Scanning electron microscope (SEM)
Atomic absorption spectrometry
Plasma emission spectroscopy
Gas chromatograph-mass spectrometry
Atomic absorption spectrometry
Purgeable organics
Non purgeable
Organics
Multi-element
Multi-element
Particular sizing and
composition
Multi-element
Multi-element
Organic
Multi-element
Species collected included:
Maianthemum dilatatum
beadruby
bracken
Pteridium aquilinum
Cornus canadensis bunchberry dogwood
queens cup
huckleberry
Clintonia uniflora
Vaccinium parvif olium
High Altitude Quinault
Species collected included:
1. Moss species
2. Polystichum species
3. Berberis nervosa
4. Gaultheria shallon
5. Rhododendron species
High Altitude Dosewallips
moss
fern
Oregon grape
salal
rhododendron
Located 50 m north of the Enchanted Valley Trail,
approximately 1 800 m northeast of the Chalet
shelter, this area was in a vegetative zone
characterized and dominated by western hemlock
Tsuga heterophylla. Associated species include
Pseudotsuga menziesii and Thuja plicata. Trees
such as Alnus rubra, Acer macrophyllum and
Castanopsis chrysophylla are widespread.
Soils are moderately deep and of medium acidity.
Organic matter is high in some areas.
Species collected included:
1.
2.
3.
4.
5.
Vaccinium alaskaense
Tiarella trifoliata
Rubus spectabilis
Lobar! urn oreganum
Moss species
Alaska huckleberry
foam flower
blackberry
lichen
moss
Low Altitude Dosewallips
This sampling area was located about 2 500 m west
of the Dosewallips trailhead on the west fork of
the Dosewallips River. It was dominated by western
hemlock and Pseudotsuga menziesii the subdominant
species, and the general aspect was similar to
that found at the high Quinault sampling site.
The location of this sample area is about 2 500 m
northeast of the Anderson Pass shelter on the west
fork of the Dosewallips River. The dominant
species and the vegetative aspect were similar to
that found at the high altitude Hoh sampling site.
Species collected included:
1. Moss species
2. Abies amabilis
3. Ribes species
4. Vaccinium ovalifolium
5. Vaccinium membranaceum
ANALYTICAL TECHNIQUES
moss
silver fir
gooseberry
oval-leaf huckleberry
thin-leaved huckleberry
All samples were prepared for analysis at the U.S.
Environmental Protection Agency's Environmental
Monitoring Systems Laboratory at Las Vegas.
Detailed procedures and techniques concerning the
analytical methods used for each sample type have
been described (Wiersma et al. 1979) and (Wiersma
and Brown 1979). Table 1 shows the analytical
methods presently being utilized for analysis of
the Olympic National Park samples.
All the samples have been prepared and presently
are being analyzed for selected contaminants.
(See Wiersma and Brown 1979).
81
CONCLUSIONS
The collection and assessment of biological and
physical media from selected areas of the Olympic
National Park will be used to develop a multimedia
pollutant monitoring system. The techniques used,
associated with long-term monitoring data, will
serve to identify baseline levels, identify
pollutant concentrations, determine trends, and
define the physical and biological responses to
synthetic contaminants.
LITERATURE CITED
Ad Hoc Task Force on GNEM.
1970. A global network for environmental
monitoring. A report to the Executive
Committee, U.S. National Committee for the
Intern. Biolog. Progr.
Morgan, G. B. , G. B. Wiersma, and D. S. Barth.
1979. Monitoring on Biosphere Reserves for
regional background levels of pollutants. In
Proc. of the United States-Union of Soviet
Socialist Republics, Symp. on Bios. Res. ,
Moscow, USSR, May 1976. USDA For. Serv. Gen.
Tech. Rep. PNW-82, p. 90-104. Pac. Northwest
For. and Range Exp. Stn. , Portland, Oreg.
O'Brien, B. J.
1979. The exposure commitment method with
application to exposure of man to lead
pollution. MARC Rep. 13. Monit. and Assess.
Res. Cent. Chelsea Coll., Univ. London.
Schuck, E. A., and G. B. Morgan.
1975. Design of pollutant oriented integrated
monitoring systems.. Intern. Conf . on Environ.
Sensing and Assessment. Pap. No. 20-6.
Barry, P. S.
1979. An introduction to the exposure com-
mitment concept with reference to environmental
mercury. MARC Rep. 12. Monit. and Assess. Res.
Cent. Chelsea Coll., Univ. London.
Wiersma, G. B.
1979. Kinetic and exposure commitment analyses
of lead behavior in a Biosphere Reserve. MARC
Rep. 15. Monit. and Assess. Res. Cent. Chelsea
Coll. , Univ. London.
Brown, K. W. , G. B. Wiersma, and C. W. Frank.
1979. A battery-operated air sampler for remote
areas. EPA-600/ 4-79-071. U.S. Environ. Prot.
Agency, Las Vegas, Nevada.
Franklin, Jerry F.
1977. The Biosphere Reserve program in the
United States. Science 195:262-267.
Franklin, Jerry F. , and C. T. Dyrness.
1973. Natural vegetation of Oregon and
Washington. USDA For. Serv. Gen. Tech.
Rep. PNW-8. 417 p.
Wiersma, G. B. , and K. W. Brown.
1979. Recommended pollutant monitoring system
for Biosphere Reserves. Proc. Second Conf. on
Sci. Res. in the National Parks. San Francisco.
Nov. 26-30.
Wiersma, G. B. , K. W. Brown, R. Herrmann,
C. Taylor, and J. Pope.
1979. Great Smoky Mountains preliminary study
for Biosphere Reserve pollutant monitoring.
EPA-600/4-79-072. U.S. Environ. Protec. Agency,
Las Vegas, Nevada.
82
Research/Management Prescribed Burning at Lava Beds National Monument
Robert E. Martin, Craig M. Olson, and James Sleznick, Jr.
ABSTRACT
Prescribed burning at Lava Beds National Monument has been developed for research and management. Research
objectives have been to develop burning techniques and prescriptions, evaluate effects, and develop cost
data. Management objectives have been to restore the historic vegetation sequence, and reduce accumulated
fuels. Twenty-seven prescribed burns have been conducted, ranging in size from 0.1 to 570 hectares (ha)
for a total of 1 088 ha or 5.8 percent of the Monument area. Prescriptions and techniques have been devel-
oped for burning all vegetation types on the Monument, primarily shrub-grass and pine-shrub types.
Robert E. Martin and Craig M. Olson, research
foresters, U.S. Department of Agriculture, Forest
Service, Pacific Northwest Forest and Range
Experiment Station, Silviculture Laboratory, Bend,
Oregon; and James Sleznick, Jr., superintendent,
Lava Beds National Monument, Tulelake, California.
83
INTRODUCTION
Lava Beds National Monument has received extensive
attention since 1973 concerning prescribed burning
and fire effects on flora and fauna. In this pa-
per, we summarize the prescribed burning portion
of the program. Prescriptions, techniques, area
descriptions, general effects, and costs of pre-
scribed burning in comparison to wildfires are used
to illustrate both the research and management as-
pects of the program. Specific effects are covered
in other papers.
The Monument lies in extreme northern California
about 80 km southeast of Klamath Falls, Oregon. A
modified maritime climate prevails with the Cascade
Range to the west providing a distinct influence on
the climate. Summers are dry, with only 9 cm of
the annual 33.7 cm precipitation falling in May
through August. Elevation ranges from about 1 230 m
(4,030 ft) in the big sagebrush, Artemisia tri-
dentata, -grass types at the north end of the Monu-
ment to 1 600 m (5,250 ft) in the ponderosa pine,
Pinus ponderosa , -ant elope bitterbrush, Purshia tri-
dentata, types in the south end of the Monument.
The north slopes of buttes, which rise to 1 670 m
(5,480 ft), often are covered by a ponderosa
pine-greenleaf manzanita, Arctostaphylos patula,
-snowbrush ceanothus, Ceanothus velutinus, com-
munity.
Fires, started
quent in most
the advent of
fires occurred
from examinati
of downed logs
perieuced an a
per million ac
Until 1950, mu
Monument with
every 5 years
by Indians o
parts of the
successful fi
every 5 to 1
on of 76 fire
Since 1933
verage of abo
res per year
ch of the nor
flashy shrub-
r lightning, were f re-
Monument area before
re control. Prehistoric
5 years, as determined
-scarred cross-sections
the Monument has ex-
ut 40 lightning fires
(Martin et al. 1977).
thern two-thirds of the
grass fuels, burned
Fire control, in connection with a catastrophic
western pine beetle, Dendroctonus brevicomis, out-
break in the late 1920' s and heavy grazing around
the turn of the century has done much to modify
the vegetation and allow heavy fuel accumulation.
Heavy grazing and fire exclusion contributed to
decline of native bunchgrasses, enhanced invasion
of cheatgrass, Bromus tectorum, and other exotic
plants, and allowed increased numbers of big sage-
brush, gray rabbitbrush, Chrysothamnus nauseosus,
and western juniper, Juniperus occidentalis. At
middle elevations, curlleaf mountainmahogany , Cer-
cocarpus ledif olius, and bitterbrush have prolif-
erated from fire exclusion following loss of pine.
These stands are now old and decadent and form an
explosive fuel complex as evidenced by the 1973
Lava Wildfire. At higher elevations, the pine-
bi tterbrush-manzanita complex, in some places
being invaded by white fir, Abies concolor , can
now be lost entirely in a wildfire. The 1977
Scarface-Mud Wildfire (36 000 ha) and 1978 Twins
Wildfire (10 000 ha) burned nearby in types simi-
lar to Lava Beds fuel types. These fires attained
large sizes even though suppression efforts were
vigorous, the terrain is very accessible, and no
extreme weather conditions prevailed. These fires
indicate a potential for catastrophic fire in the
Monument.
OBJECTIVES OF THE RESEARCH /MANAGEMENT PRESCRIBED
BURNING PROGRAM
The research/management prescribed burning program
has several objectives. These may conveniently be
separated into two groups, though there is overlap
between them.
Research
(1) Develop burning techniques and prescriptions
to achieve objectives in fuel and vegetation
management.
(2) Record effects of fire under differing condi-
tions on flora, fauna, and fuels.
(3) Develop cost data.
Management
(1) Restore "historic" vegetation sequence.
(2) Reduce accumulated fuels.
(3) Train personnel in fire use.
(4) Develop fire management plans and programs.
The interdependency of the objectives may or may
not be apparent. The prescription and technique
development, however, has often been conducted on
management-sized units. Thus, cost data can be
obtained, along with flora and fuel changes. Mon-
ument personnel are trained on the burns in re-
ducing fuels and the beginning steps of restoring
the historic vegetation sequence that have been
taken. The historic vegetation sequence is con-
sidered to be that which would have existed had we
not interfered following the 1873 Modoc Indian War,
which occurred on the northern end of the Monument.
The supposed vegetation has been constructed from
photographs, live and dead vegetation, historic
accounts, and fire history. Since most burns have
been planned in strategic locations, they fit into
overall fire management planning. We have been
able to move toward accomplishment of the multiple
objectives by combining the efforts of two small
staffs at the Bend Silviculture Laboratory and
Lava Beds National Monument, with the assistance
of personnel from the Doublehead Ranger District
of the Modoc National Forest.
Our main discussions will center on prescriptions
and techniques for the major vegetation complexes
of the Monument, the costs for burning these areas,
and how these fit into fire and vegetation manage-
ment of the Monument. Effects of fire on flora
and fauna are covered in several other papers in
this report. Johnson and Smathers (1976) discussed
historic vegetation on the Monument and the role
of grazing, beetles, and fire control in changing
it. Martin and Johnson (1979) discuss stages of
the prescribed burn research program, including
fire history, succession under varying fire re-
gimes, and general management planning. Other
papers relating to fire effects in the Lava Beds
cover small mammals (Frenzel 1978, Frenzel et al.
1979), Columbia sharptail grouse, Pedioecetes
phasianellus, (Starkey and Schnoes 1979), deer,
Odocoileus hemionus , (Schnoes 1977), bobcats, Lynx
rufus, (Zezulak 1978), vegetation (see Olson in
this report), and bird nesting (see Tiagwad et al.
in this report). We will only allude to results
of these studies here as space will permit us to
show only the relationship to prescription and
technique development.
84
PRESCRIPTION RECOMMENDATIONS
The following prescription conditions are for the
major vegetation types of Lava Beds National Monu-
ment and were developed from our experience at the
Monument and elsewhere. When considering condi-
tions for burning, never use all the "worst" or
"best" conditions for a given burn. For example,
don't pick, the driest fuels, lowest humidities,
and highest temperatures in a situation where heavy
fuels exist in stands with low crowns. Prescrip-
tion conditions, as given, represent a range of
conditions we have found to be satisfactory. Mod-
ifications should be made to fit specific vegeta-
tion conditions or transitions from one type to
another.
Prescribed burning began in 1974 with two 0.01-ha
units burned in cheatgrass. From 1974 to 1979,
27 fires have been conducted, some as reburns to
evaluate the effect of repeated fires (table 1).
The largest unit of 570 ha (1,400 acres) was burned
in the fall of 1979. Little burning was done in
1978 because resources were limited. The units,
their size, burning conditions, and specific com-
ments pertinent to each burn are given in table 1
for the information of those who might use fire in
similar vegetation and fuel situations. These data
contributed to the prescription ranges given below.
Cheatgrass
Burning has reduced cheatgrass seed density on soil
surfaces, although apparently not enough to enable
native bunchgrasses to become reestablished rapidly
(Olson et al. 1981). Where some native bunchgrass,
particularly bottlebrush squirreltail, Sitanion
hystrix, is present, burning has appeared to in-
crease both vigor and density of this species even
though our early sampling techniques were not sen-
sitive enough to record these changes.
The flammability and horizontal fuel conductivity
of cheatgrass stands permit a relatively wide range
in prescription conditions. Prescription develop-
ment has the important objective of reduction of
cheatgrass as an undesirable exotic invader.
Season: Anytime following cure in June until fall
precipitation.
Wind: 0 to 15 km/hr.
Relative humidity: 20 to 45 percent.
Precipitation: Preferably at least 3 days follow-
ing precipitation to permit duff drying and higher
seed kill.
Temperature: 10° to 30°C.
Burning pattern: Backfire downwind line; strip
headfire to develop burnout of 10 to 30 m, depend-
ing on wind, humidity, and fuel; headfire unit;
center or ringfiring can be used under zero wind
conditions.
Fireline: Excellent fuel for use of wetline (Mar-
tin et al. 1977); where duff exists, extra caution
in use of wetline is necessary.
Burns in these types are generally conducted to
reduce shrub and increase native bunchgrass cover.
In our burning, we have sought to achieve a mosaic
of burned and unburned areas within the burn units.
If the original burn does not achieve a satisfac-
torily high burn percent, additional ignition can
be used to increase burn percent in the days fol-
lowing the main burn.
Season: Early summer through late fall will be
satisfactory; if bunchgrass recovery is a major
consideration, late fall has been most productive
(see Olson et al. in this report).
Wind: 3 to 16 km/hr most desirable; lower winds
will reduce probability of spread if fuels are
discontinuous; gusts to 25 km/hr or higher are no
problem if fire is not near fireline.
Relative humidity: Line burnout 20 to 40 percent;
general burn 15 to 40 percent; in early season
burns when grasses are not cured, humidity should
be 30 percent or less.
Precipitation: Not critical; could be only 1 or 2
days before fire.
Temperature: 10° to 30°C.
Burning pattern: Back out downwind lines, * then
strip headfire'' 30 to 80 m, depending on wind
fuels; then headfire.
Fireline. Wetline-' or handline.
Ponderosa Pine
Most ponderosa pine types have bitterbrush as an
understory. Small areas of pine within the Monu-
ment may have big sagebrush, greenleaf manzanita,
or snowbrush ceanothus as the dominant shrub.
Prescriptions would not be much different for the
other shrubs, although slightly drier conditions
may be appropriate.
-"-Set a fire that burns into the wind from the
downwind side of the fireline.
^A fire lighted at a designated distance upwind
from the burned strip resulting from the back out
fire.
-'A fire burning with the wind.
^An established or natural break in the fuel,
generally down to mineral soil.
-'Two wet strips, usually made by a pumper unit,
with a dry area between.
"A fuel break dug with hand tools.
Terminology has been taken from Glossary Terms for
Fire Management Planning, USDA For. Serv. , January
1979; and Planning for Prescribed Burning in the
Inland Northwest, by Martin and Dell, USDA For.
Serv. Gen. Tech. Rep. PNW-76; 1978.
Shrub-Grass
Dominant shrubs in these situations are primarily
sagebrush, gray rabbitbrush, or bitterbrush.
Grasses may be quite varied but are generally a
mix of bunchgrasses interspersed with cheatgrass.
85
Table 1 — Summary table of prescribed burns conducted at Lava Beds National Monument, 1974 to 1979
Prescribed burn
Date burned
Vegetative types
Size
Temperature
Humidity
ha
°C
°F
Percent
Cheatgrass
West Wildlife Overlook. 1
West Wildlife Overlook 2
Gillem Cheatgrass
Lava Overlook 1
Lava Overlook 2
Lava Overlook 3
Homestead Flow
Upper Ice Cave 1
Upper Ice Cave 2
Upper Ice Cave 3
Hovey Point
Fleener Chimneys
Little Crescent
Caldwell Loops-Upper
Caldwell Loops-Lower
Headquarters
East Caldwell
Gillem Camp 1
Gillem Camp 2
Gillem Camp 3
East Wildlife Overlook
Sheep Enclosure
Black Lava Flow
Captain Jack's Bridge
Schonchin
6/ — /74 Cheatgrass
6/25/75 Sagebrush-grass
6/24/76 Cheatgrass
6/26/75 Cheatgrass
6/26/75 Sagebrush-bunchgrass
8/22/75 Sagebrush-bunchgrass
11/2/75 Sagebrush-bunchgrass
6/26/75 Sagebrush-bunchgrass
6/27/75 Ponderosa pine-bitterbrush
8/22/75 Ponderosa pine-bitterbrush
11/3/75 Ponderosa pine-bitterbrush
6/23/76 Sagebrush-grass
6/2/76 Juniper-sage brush- bunchgrass
9/20-21/76 Juniper-mountainmahogany-
bitterbrush
9/21/76 Ponderosa pine-bitterbrush
9/21/76 Ponderosa pine-bitterbrush
7/7-8/77 Juniper-mount a inmahogany-
bitterbrush
5/19/77 Ponderosa pine-bitterbrush
7/7/77 Juniper-sagebrush-grass
7/5/78 Cheatgrass
6/28/79 Cheatgrass
9/19/78 Sagebrush-grass
7/5/78 Sagebrush-grass
6/6/79 Ponderosa pine-bitterbrush
7/2-3/79 Junlper-sagebrush-bunchgrass 40.0
9/11-12/79 Juniper-sagebrush-bunchgrass
Juniper-mount a inmahogany-
bitterbrush 570.0
0.1
7.3
14-18
58-64
33-52
2.0
20-21
69-70
34-35
0.8
17
63
42
0.15
21
69
30
0.15
21-24
69-76
33-46
0.14
15
59
47
12.0
21-22
69-71
23-31
0.7
6-16
42-61
27-62
0.7
24-26
76-79
21-25
2.0
14-19
57-67
28-50
70.0
17-22
63-71
27-42
80.0
20-24
68-76
25-27
20-23
68-73
16-28
15-22
59-72
23-48
15-24
59-75
17-23
20.0
21-28
70-82
15-25
27.0
7-28
44-70
25-80
23.0
23-26 73-78
21-24 69-75
10-15 50-85
26-30
1.5
18-27
65-81
31-47
0.6
19-23
66-74
43-48
12.0
14-19
57-67
21-38
2.0
18-23
65-74
34-47
16.0
4-12
40-53
45-60
29-39
18-43
86
Fuel moistures
Time lag class
Live
Itter
Duff
1
10
100
1000
16
9
7
4
—
7
6
—
—
11
13
—
7
7
—
—
9
—
14
28
—
—
12
—
10
11
—
—
12
—
16
—
—
—
15
115
14
14
—
—
10
34
11
14
20
228
14
15
10
10
7
9
5
10
7
9
13
9
12
13
12
13
9
12
13
12
Foliage Twigs Grass
Wind
Percent
burn
Comments
133
89
129
159
123
74
74
74
5 4 116
Destroyed in oven
7 7 194
134
123
22
km/hr
181
137
6-16
105
20
6-8
128
5-10
154
101
9-14
136
77
0-10
133
28
6-12
166
109
6-9
3-11
2-6
calm
36
5-12, Gl
40
13-19
70
25
0-16
66
0-8, G16
66
0-8, G16
72
48
0-16
0-11
0-11
80
100
100
100
100
100
90
100
50
80
70
20
40
80
30
33
6-11
90
7
3-12
80
9
0-16
90
3-18
70
0-12
90
§lt an ion hystrix 98-
percent M.C.
21
0-11
1-2 cm rain 8/18/75
Poor coverage of plot
Line burnout 40- to
50-percent RH
2-stage, resumed after
Caldwell
2-stage
Stopped burning at 17-
percent RH
3-stage burn; 1.1 cm
rain in 3 days prior
Needed lower humidity
and higher grass cure
Reburned cheatgrass
10/10/79 (2 burns)
Prescribed burning
workshop
Lines burned out at
39 to 52-percent RH
111
18
0-10
45
87
Sapling Stands
EFFECTS ON FUELS AND VEGETATION
For sapling stands up to 10 cm (4
breast height (d.b.h.). burning p
erally must allow for cool weathe
winds, and only partial consumpti
Where extensive shrubs exist unde
may be necessary to burn in stage
stage would burn the needle drape
diately under trees, and the seco
burn in the needle litter under t
adjacent shrub areas exist around
with no needle component, a third
be needed. The entire sequence c
met in 2 days of drying following
spring or extensive rain in the a
Sapling-Shrub
in) diameter at
rescriptions gen-
r, strong steady
on of large fuels,
r these trees, it
s. The first
in shrubs imme-
nd stage would
he trees. Where
the trees but
, drier stage may
ould generally be
a light rain in
utumn.
Season: Spring, early summer, or autumn when large
fuels are wet.
Wind: 2 to 16 km/hr in stand.
Relative humidity: 60 to 85 percent.
Precipitation: Light rain 1 or 2 days previous in
April to June; generally 2.5 to 4 cm precipitation
in autumn.
Temperature: 7° to 20°C.
Burning pattern: Generally this will be done by
lighting individual shrubs or shrub patches so the
fire backs through them; general pattern of light-
ing is not crucial, but the easiest pattern would
be a spot headfiring pattern, even though indi-
vidual fires will seldom spread. The second stage
of this burning would be conducted under conditions
for burning sapling-pine litter below.
Fireline: In pines, wetline can be used with re-
tardants or without only where no thick duff
exists.
Sapling-Pine Litter
Season: Spring, early summer, autumn; 2.5 to 4 cm
precipitation in autumn will reduce tree damage by
reducing fuel consumption.
Wind: 1 to 16 km/hr in stand.
Relative humidity: 30 to 60 percent.
Precipitation: Precipitation 2 to 7 days before
burning can help reduce damage where heavy duff is
present.
Temperature: 5° to 25°C.
Burning pattern: Back and strip headfire; adjust
flame length according to wind and temperature
using Van Wagner's scorch height curves as a guide-
line (Albini 1976).
Mature Pine-Bi t terbrush-Pine Litter
Season: Spring, early summer
ditioning by one or two burns
should present no problems.
Wind: 1 to 16 km/hr in stand
Relative humidity: 40 to 70
burn; 20 to 50 percent for su
Precipitation: In autumn, bu
ing 2. 5 cm or more of rain af
Temperature: 5° to 25°C for
for subsequent burns.
Burning patterns: Back and s
first burn, adjusting strip h
desired flame length; in cond
outline and then strip headfi
conditions permit.
, autumn; after con-
, summer burning
percent for first
bsequent burns.
rn 2 to 4 days follow-
ter September 1.
first burn; 5° to 30°C
trip headfire for
eadfire width to get
itioned stands, back
re or headfire as
The burns have had somewhat different effects on
fuels and vegetation, depending on the season and
conditions under which they were conducted. De-
tailed information on vegetation effects is given
elsewhere (see Olson et al. in this report). The
effects on fuels are supplemented by data we have
collected in similar burns elsewhere.
Cheatgrass
Burns have been conducted in early summer and au-
tumn, with sampling on fuels, measurement of cover,
and bioassay of cheatgrass and other annual seeds
before and after burning. Some areas have not
been burned in three successive years to evaluate
any cumulative effect of fire on cheatgrass cover
and seed count. Some reduction in seed count has
been achieved. More important, the fire seemed to
restore vigor to scattered bunches of bottlebrush
squirreltail, as well as encouraging new plants of
this native species. The reinvigorated bunches
apparently exert a strong allelopathic effect on
surrounding cheatgrass, but the effect appears to
diminish by the 3d year.
Prescribed burning removes almost all standing fuel
in cheatgrass stands, the remaining standing fuels
generally being uncured cheatgrass, bunchgrasses,
or mustards. The amount of duff consumed varies
depending on humidity and recent weather condi-
tions. Fire hazard is negligible until the cheat-
grass cures the following summer.
Shrub-Grass
Effects of fire on vegetation in shrub-grass types
are strongly dependent on burning and vegetation
conditions. Weather conditions at the time of the
burn may also alter the effect of fire on vegeta-
tion. There is a limited range of conditions under
which fire will spread in these types. Previous
precipitation may be important in the degree of
shrub sprouting.
Big sagebrush and mountainmahogany are readily
killed by fire and generally do not sprout. Up to
80 percent of gray rabbitbrush plants are killed
by spring and early summer fires. Up to 30 percent
of bitterbrush plants may sprout when burning is
conducted under cool, moist autumn conditions,
lesser numbers in early summer. Green rabbitbrush,
Chrysotharanus viscidif lorus, and blooming rabbit-
brush, Haplopappus bloomeri , survive burning under
a wide variety of conditions, and bitter cherry,
Prunus emarginata , sprouts prolif ically even after
very severe wildfires.
The native bunchgrasses respond well to burning
under most conditions depending on phenological
stage. Bluebunch wheatgrass, Agropyron spicatum,
Idaho fescue, Festuca idahoensis, bottlebrush
squirreltail, and giant wildrye, Elymus cinereus,
recover rapidly from burning. Others such as
Sandberg's bluegrass, Poa sandbergii, may be set
back considerably by burning. Uncured perennial
grasses seem to be most sensitive when fire is
carried by cured annuals such as cheatgrass.
88
Fuels in the shrub-grass type are reduced so that
fire hazard is very low until the next summer.
Grass and forb fuel weights are reduced from around
0.3 t/ha to around 0.1 t/ha.
Most dead shrub components, both standing and
down, are consumed during burning, and many live
components are converted to standing dead. In
sagebrush, the partly live stems with shredded,
flammable bark are often burned off, thus becoming
dead and down fuel.
Fuels before and after burning on the East Caldwell
unit in pine-bitterbrush are given in table 3. The
data were from 24 samples of each, and the reduc-
tions of litter and material > 7.62-cm diameter
were significant at the 99-percent level of confi-
dence. The increase in 0 to 0. 63-cm down and dead
fuel class was caused by burning off stems which
fell to the ground after the fire had passed. The
same phenomenon also accounted for an apparently
lesser fuel reduction in the 0.63 to 2.54- and
2.54 to 7. 63-cm size classes.
Fuel reduction on an area basis depends on percent
shrub cover before burning. If canopy cover were
100 percent before burning, total fuel loads for
sagebrush and bitterbrush are given in table 2.
Fuel load for partial canopy cover can be calcula-
ted by multiplying by the fraction of area covered
by shrub canopy. Most of the burns in the shrub-
grass types have been designed to cover 50 to 70
percent of the area within the firelines, leaving
a mosaic of vegetation. Fuel reduction and vege-
tation effects are measured as burned or unburned
area to give more accurate indication of fire
impact. The mosaics enhance wildlife habitat,
improve esthetics, and approach the natural fire
effects.
FIRE MANAGEMENT PLANNING
Prescribed burns at Lava Beds National Monument
have been arranged in patterns to isolate parts of
the Monument from others (fig. 1). By conducting
burns in these patterns, the fire hazard will not
be eliminated but the fuel complex will be more
manageable. In shrub-grass fuels, prescribed burn-
ing has removed heavy accumulations of dead woody
material, generally converting the vegetation to a
grass-forb complex for several years. Although
the grass-forb complex may allow for more rapid
fire spread under some conditions, fire intensity
and spotting potential will be reduced drastically,
thus making fire control easier.
Pine-Bitterbrush
Season and precipitation prior to burning are very
important in the effects of fire on vegetation in
this type. Effects on shrub and herb vegetation
are similar to that in shrub-grass types. Ponder-
osa pine is quite resistant to fire except as a
seedling, and even larger seedlings can withstand
a very light fire. The tree is killed by crown
scorch or by girdling near the base. At Lava
Beds, we have killed trees by girdling only when
the trees were very small (less than 5-cm diameter
at the base) and growing in heavy duff, or when
logs or stumps were consumed next to larger trees.
Most tree mortality has resulted from crown scorch
(or consumption). In the August 1975 Upper Ice
Cave burn, 29 percent of the pine were killed. In
no other burns at Lava Beds, has tree kill exceeded
13 percent, and these were the smaller trees in the
stands (see Olson et al. in this report).
Table 2 — Live and dead fuel loads for big sagebrush
and antelope bitterbrush with 100-percent shrub
canopy cover
Fuel diameter by size
classes, (centimeters)
Foliage Total
0-0.63 0.63-2.54 2.54-7.
63
- tonnes per hectare - -
Sagebrush:
Live
1.4
1.7 2.7 1.9
7.7
Dead
—
3.6 1.9 0.5
6.0
Bitterbrush:
Live
0.2
1.8 1.1 0.2
3.3
Dead
"
4.5 3.1 0.4
8.0
Table 3 — Fuels before and after burning on the East Caldwell unit in pine-bitterbrush
Fuel size class (centimeters)
Preburn N = 24
Postburn N = 24
0-0.63 0.63-2.54 2.54-7.63 7.63+ Litter
0.0.63 0.63-2.54 2.54-7.63 7.63+ Litter
tonnes per hectare
0.272
0.337
1.1555
1.239
2.502
3.063
59.7
87.0
9.63
9.15
0.486
0.619
0.682
0.604
2.096
2.001
1.48
5.51
6.01
7.33
89
Figure 1. — Prescribed burns and wildfires (black),
planned burns — fuel modification areas (cross
hatching), and lava flows (dash) on Lava Beds
National Monument.
Table 4 — Costs of prescribed burning and wildfires
recorded at Lava Beds National Monument
Area
Total
Cost per
Year
(hectares)
cost
hectare
PRESCRIBED FIRES1
Hovey Point
1976
170
$ 325
i 1.91
East Caldwell
1977
27
1,389
51.44
Headquarters
1977
120
3,200
26.67
Schonchin
1979
570
2,068
3.63
WILDFIRES
Bighorn Fire
1973
121
$ 39,600
$ 326
Lava Fire
1973
207
132,000
638
Cougar Fire (Modoc
National Forest)
1977
1A6
125,560
860
Strike Fire
1977
117
149,400
1,277
^To obtain acres from hectares, multiply by 2.47; for
cost per acre, divide cost per hectare by 2.47.
In ponderosa pine stands, prescribed burning re-
duced fuels, making fire control relatively easy
under most wildfire conditions. Potential high
rate of spread and extreme fire intensity were
reduced, at the same time making the pine stand
quite resistant to fire.
COSTS OF PRESCRIBED BURNING
Costs of prescribed burning will be an important
factor governing how much burning is done. The
Monument has recorded costs on most of the large,
management-sized prescribed burns and wildfires
since 1976. Costs are given in table 4.
Costs of prescribed burns included the cost of
preparing firelines, conducting the burns, and
mop-up; not included were the costs of planning
and data gathering. Planning of burns was closely
tied in with general fire management planning and
the two were difficult to distinguish. Sampling
of fuels and vegetation, and fire documentation
were considered a research function and were not
included for that reason. Size of burn, type of
vegetation, method and amount of line preparation,
and weather variables were major factors affecting
the cost of prescribed burning.
Hovey Point and Schonchin burns were in shrub-grass
types. Firelines on Hovey Point were all roads or
were prepared by the wetline method. Line prepara-
tion was begun by five people at 0730. By 0930,
backfiring began, and the unit was burned out by
1300, with a maximum of 12 people involved. Flame
lengths varied from 0.5 to 10 m, but were 2 to 5 m
throughout most of the burn. Highest rates of
spread were not over 2 km/hr.
The Schonchin unit required about 120 m of line
preparation prior to burning. Line burnout began
at 1900 one evening, continued until 0330, and re-
sumed at 0800. The unit burn began about 1100 and
was completed about 1600 with 40 percent of the
unit burned. Flame lengths varied from 0.5 m in
grass areas to 10 or 15 m in sagebrush and moun-
tainmahogany thickets, with occasional western
juniper ignitions reaching to about 20 or 25 m.
Rates of spread were estimated to range up to
5 km/hr and depended primarily on wind and fuel
conditions.
The Headquarters prescribed burn was relatively ex-
pensive for three major reasons. First, a line was
cut almost one-half km through mountainmahogany-
bitterbrush fuels. Second, extensive burnout was
conducted on the southeast side of the burn to iso-
late Monument buildings and dangerous fuel concen-
trations. Third, two attempts to ignite the unit
were unsuccessful. Unit burnout began with strip
headfiring, but an alternative plan for centerfiring
was used when the winds subsided. Flame lengths
ranged from 2 to 10 m in sagebrush-bitterbrush
fuels, 5 to 10 m in mountainmahogany, with flames
to 20 m when western junipers ignited. Rates of
spread up to 3 km/hr were estimated.
90
The East Caldwell unit required 200 m of line
preparation through mountainmahogany-bitterbrush
fuels. The unit contained scattered ponderosa
pine and pine thickets. Most pine were saplings
or poles with bitterbrush underneath. The unit
was burned in three stages: (1) bitterbrush
under pines were burned the morning after a light
rain; (2) that afternoon the litter under the pine
stands was burned; and (3) in the next 2 days, the
rest of the unit was burned.
In contrast to the prescribed burns, recent wild-
fires in the area cost from $300 to $1,300 per ha
to control, not including damages, if any. The
Bighorn and Lava wildfires were both started by
the same lightning storm in July 1977. Vegetation
in the Bighorn fire was sagebrush-grass, and the
fuel load was probably somewhat less, on the aver-
age, than that of the Hovey Point and Schonchin
prescribed burns. Fuel load was definitely much
less than that in the higher, wetter end of the
Schonchin burn. Ratio of costs for wildfire con-
trol to prescribed burning in these vegetation
types are 171:1 and 90:1.
SUMMARY
Research and management objectives have been met
through the prescribed burning program at Lava Beds
National Monument. Burning prescriptions and tech-
niques have been developed and effects on vegeta-
tion measured, although more information is needed
concerning effects of fire on flora and fauna. The
larger burns fit into plans for fire management
planning and have provided cost data for planning
future burns.
LITERATURE CITED
Albini, Frank A.
1976. Estimating wildfire behavior and effects.
USDA For. Serv. Gen. Tech. Rep. INT-30, 92 p.
Intermt. For. and Range Exp. Stn. , Ogden, Utah.
Frenzel, Richard W.
1978. The effects of prescribed burning on small
mammal communities in Lava Beds National Monu-
ment. M.S. thesis, 79 p. Oreg. State Univ.,
Corvallis.
The Lava fire burned in fuel types very similar to
those of the adjacent 1977 Headquarters prescribed
burn. Maximum number of people on the wildfire was
375 compared to 11 on the prescribed burn. Ratio
of wildfire to prescribed burn costs per hectare
are 24:1.
Frenzel, R. W. , E. E. Starkey, and H. C. Black.
1979. Effects of prescribed burning on small
mammal communities in Lava Beds National Monu-
ment, California. Proc. of the First Conf. on
Scientific Res. in the Natl. Parks. NPS Trans,
and Proc. Series 5(1) :287-292.
The Cougar and Strike fires occurred in mixed shrub
and pine-shrub types most similar to the East Cald-
well burn, which was small enough that it suffered
from economies of scale. Still the two wildfires
cost 17 and 25 times as much per hectare as the
prescribed burn. The wildfires resulted in almost
total loss of the sapling-pole-sawtimber stands,
most of it unsalvageable, whereas only 13 percent
of the trees were killed in the prescribed burn.
Presently, about 1 088 ha or 5.8 percent of the
Monument's 18 720 ha have been prescribed burned.
Three-quarters of the Monument, or 14 000 ha, could
be burned at the cost per hectare of the Schonchin
unit cost or less, or approximately $51,000. By
considering the economy of burning large units, and
eliminating lava areas, the remainder could prob-
ably be burned for $15 per ha or $70,000 in today's
money. Thus, a program to reintroduce fire to the
entire Monument could be conducted for less than
the 1973 Lava wildfire, even disregarding the de-
creased value of the dollar.
Johnson, Arlen H. , and Garrett A. Smathers.
1976. Fire history and ecology in Lava Beds Na-
tional Monument. Proc. Annu. Tall Timbers Ecol.
Conf. 15:102-115.
Martin, Robert E. , Stuart E. Coleman, and Arlen H.
Johnson.
1977. Wetline technique for prescribed burning
firelines in rangeland. USDA For. Serv. Res.
Note PNW-292, 6 p. Pac. Northwest For. and
Range Exp. Stn. , Portland, Oreg.
Martin, Robert E. , and Arlen H. Johnson.
1979. Fire management of Lava Beds National Mon-
ument. Proc. of the First Conf. on Scientific
Res. in the Natl. Parks. NPS Trans, and Proc.
Series 5(2) :1209-1217.
Schnoes, Roger S.
1977. The effects of prescribed burning on mule
deer wintering at Lava Beds National Monument.
M.S. thesis, 70 p. Oreg. State Univ., Corvallis.
Prescribed fires cannot eliminate the chance of
wildfire and subsequent costs; but they do reduce
hazard and increase the ease of control, while a-
voiding catastrophic wildfire. The initial costs
of reconditioning fuels in the Monument will prob-
ably be much greater than costs of a maintenance
prescribed burning program. Some areas of the
Monument have accumulated heavy fuel loadings and
may require two or three prescribed fires to reduce
loadings to a manageable level. But even this
would be less expensive on a per hectare basis than
a catastrophic wildfire. Refined prescriptions,
trained personnel, and manageable fuel complexes
will enable further reductions in prescribed fire
costs while approaching objectives in vegetation
management.
Starkey, E. E. , and R. A. Schnoes.
1979. The Columbian sharptailed grouse: with
special reference to their potential reintroduc-
tion to Lava Beds National Monument. Proc. of
the First Conf. on Scientific Res. in the Natl.
Parks. NPS Trans, and Proc. Series 5(1) .-497-500.
Zezulak, David S.
1978. Northeastern bobcat study. Draft Final
Rep. Calif. Fish and Game Dept. June 1978.
91
Effects of Prescribed Fires on Vegetation in Lava Beds National Monument
Craig M. Olson, Arlen H. Johnson, and Robert E. Martin
ABSTRACT
Prescribed fire is being used in an attempt to restore the historic vegetation sequence of Lava Beds
National Monument where, prior to settlement, fires occurred every 5 to 15 years. Results of the burning
will be used to develop fire plans for the Monument.
Craig M. Olson and Robert E. Martin, research
foresters, U.S. Department of Agriculture, Forest
Service, Pacific Northwest Forest and Range
Experiment Station, Silviculture Laboratory, Bend,
Oregon; Arlen H. Johnson, faculty of forestry,
University of British Columbia, Vancouver, B.C.,
Canada.
92
INTRODUCTION
Fire plays an important role in determining vegeta-
tion structure and composition throughout most of
the Western United States. Successional changes
following fire are determined by the surviving
species and proximity of potential invading
species. In Lava Beds National Monument, Calif-
ornia, dramatic changes have occurred in the plant
associations with the coming of settlers and sub-
sequent alteration of fire frequency, intensive
livestock grazing, and introduction of exotic
plant species. In 1974, research was begun in
Lava Beds to evaluate the effects of fire on the
different plant associations, to identify probable
presettlement plant associations, and to recommend
practices that could help reestablish presettle-
ment plant associations.
Prescribed burning has been conducted on 22 units
on the Monument from 1974 to 1979. Some required
more than 1 day to complete burning while others
were burned in less than 1 hour. Some were burned
only once, while others were burned 2 or 3 years
in succession. Research on these units was
designed to: 1) evaluate the effects of various
environmental conditions on fire behavior; 2)
establish survival rate of grasses, shrubs, and
trees; and 3) establish successional relationships
following fire of the principal competitors within
different plant associations. This paper dis-
cusses the effects of prescribed burning under
varying conditions on the major plant associations
through the first few years following fire.
Because replication of prescribed burns is diffi-
cult, if not impossible, results reported in this
paper represent individual case studies. Infer-
ences can be made though, when several burn units
are located within similar plant associations.
STUDY AREA
Lava Beds National Monument is located in northeast
California along the north slope of the Medicine
Lake Highlands, a shield volcano. Elevation
ranges from 1 200 meters (m) along Tulelake on the
north to 1 700 m in the south. Annual precipita-
tion averages 33.7 centimeters (cm), most of it
falling from October to May. July daytime high
temperatures average 27° celsius (C) while the
daily mean low temperature for January is -6°C.
The soils in the Monument are derived from pyro-
clastic debris and basalt. They are relatively
young with little horizon development. A stone
pavement overlies shallow gravelly sand to sandy
loam. There are many basalt outcrops and recent
lava flows (<2,000 years old).
Vegetation associations at the lower elevations
are often dominated by big sagebrush, Artemisia
tridentata, or gray rabbi tbrush, Chrysothamnus
nauseosus, with cheatgrass, Bromus tectorum,
bluebunch wheatgrass, Agropyron spicatum, Thurber's
needlegrass, Stipa thurberiana, and bottlebrush
squirreltail, Sitanion hystrix, as the principal
grasses. Some low lying areas are dominated by
cheatgrass, bottlebrush squirreltail, giant
wildrye, Elymus cinereus, and needle and thread,
Stipa comata. Much of the vegetation on the
higher south end of the Monument is dominated by
ponderosa pine, Pinus ponderosa, with an antelope
bitterbrush, Purshia tridentata, understory. The
transition plant association between these two
areas is dominated by western juniper, Juniperus
occidentalis, curlleaf mountainmahogany , Cerco-
carpus ledifolius , big sagebrush, and bitterbrush
with Idaho fescue, Festuca idahoensis, bluebunch
wheatgrass, western needlegrass, Stipa occiden-
talis, and cheatgrass in the understory.
During 1975, 1976, and 1977, 12 units in the four
major associations were prescription burned. Per-
manent 10-m transects were randomly located within
and adjacent to each unit prior to burning. Units
dominated by cheatgrass were sampled using the
point contact method with 200 points per 10-m
transect (Johnston 1957). All other units were
sampled using a line intercept method (Canfield
1941) recording only the crown intercept of peren-
nial shrubs and grasses. All transects were
sampled each year at approximately the same date
as the preburn inventory.
On cheatgrass-dominated units burned in 1975 and
1977, litter and top soil were collected within
0.093-m^ quadrats before and after burning to
determine the number of viable cheatgrass seeds
surviving the fire. The samples were replanted in
a greenhouse and the number of germinants recorded.
A t-test was used to compare preburn and postburn
viable seed density.
Circular and rectangular macroplots were estab-
lished on certain burns to count the number of
sprouting and surviving shrubs by species. All
shrubs greater than 20 cm but less than 2 m in
height were recorded by species before and after
burning.
To determine if burning was significantly changing
the diameter size-class distribution of ponderosa
pine trees, the diameters of all the trees within
the units burned in 1975 and of all the trees
within twelve 0.03-ha macroplots burned in 1977
were recorded before and after burning. A t-test
was used to compare average diameter of surviving
trees with average diameter of trees killed.
Complete weather and fuel-moisture-content data
for each burn are listed in this publication (see
Martin et al. ).
RESULTS
To simplify discussion, four major plant associa-
tions are considered here. They include ponderosa
pine/bitterbrush, mountainmahogany-big sagebrush,
shrub/grass, and cheatgrass associations.
Ponderosa Pine/Bitterbrush Association
Four prescribed burns were conducte
ponderosa pine/bitterbrush associat
1975 and one in 1977. The units bu
were dominated by old-growth ponder
an average d.b.h. of 37.5 cm and a
of 82 stems per ha. The unit burne
dominated by sapling and pole-sized
with an average d.b.h. of 21.8 cm,
height of 12 m, and stand density o
ha. Preburn bitterbrush crown cove
sects ranged from 23 to 32 percent
d within
ions, three in
rned in 1975
osa pine with
stand density
d in 1977 was
ponderosa pine
an average
f 165 stems per
r along tran-
(table 1).
93
c
<u
E
-
C
o
£
n cm
in d
>aroo>iTol*©r*«aim©
M-JOOOO-JNN
r-1 00
*43 O
OOu^oOOfHCsir-l
O m O <n N H
a.
in
U-.
■
O
4-1
u
u
CV
u
LO
Jj
c
ra
4-1
«
X
4-1
c
>-
01
t->
■0
<D
>,
u
-C
4-1
c
m
o
•^
.c
c
oi
m
>-
4-1
T
0-
w
•^■CNOOOOOO-3-OCN
vO r-I
O (N ^H O O r^ H
'ho in d
>3- in as o
nHOOOOO-norg o
rHinfHi-HP^^HO\f-l (NO
m ' ' ' ri ' ' ' +
tfl
<*
CO
•u
C
0)
TJ
*n
0)
U
■u
9
CO
en
r,
■rt
tJ
u
<D
■u
V-
u
<
S
CO
CO
(11
3
•H
4J
u
01
(fl
(0
c
m
c
(J
oj
u
0
c
»a.c
oi
u
CD
m
•n
«
Tl
nH
R
c
(;
a
o
(j
u
m
o
m
>. o
3
a. 3
ra
C
o
*j
a.
2
a-
V-t
C/l
^
11
on
O cOrimoHNNOwnn
H ^' rl d O O O H O O O
s£>rom>x>eNroas<Hi-(vooo
inmr-ivooooeNmo
O ^£> o o o <r O
r-1
~o
o
o
o
o
o
o
o
0SI
o
co rH
a* m
CN -» -»
<N1
CM
PI
ti
(0
4-1
3
a
CU
a
O
I
u)
00
i-t
B
(0
4J
IH
n>
o
X
c
e
■H
4-)
CO
rH
4-1
o
ra
a
01
•a
a
ra
c
ro
c
u
3
■o
01
3
u
0)
4J
<u
£>
4-J
v-i
a
a
a
(0
n
iH
•o
tH
o
e
-o
(0
0)
M
■h
4J
B
W4
4-1
s
4-1
3
a.
^
01
(0
>,
XI
(M
•H
4-1
(0
3
4-4
«
3
a
■
C
en
n)
•o
M
3
J=
u
h
3
B
a)
4-1
C
4-1
3
•o
X
i_i
Q.
3
C
aj
c
OJ
>-.
m
c
c
e
a.
B
c
(J
'Jl
a.
c
-C
01
B
01
a
e
01
-rH
01
4->
V
3
14
3
o
li
(J
o
•
«
o
i-i
■o
o
>
■o
o
<J
Crt
T)
3
■a
0J
nj
C
u.
a
4-1
o
L>
ft
•H
o
o
•H
•r-l
V-
u
O
>>
(j
01
SI
o
c
CO
U
--(
(0
k
U)
B
h
11
u,
01
O
00
a.
3
>^
(0
X
in
<0
a.
4-1
-o
CD
u
-o
0)
1)
>
>J
u
o
o
4-1
jC
a.
0)
u
a.
4-1
4-J
u
u
cn
•H
u
3
a,
C/J
4j
c/>
oJ
SB
2
oJ
■
2
u
<
Bl.'
<*'
<
c^'
0J
u
00
<
«
4-1
GO
(0
(J
ol<|a.|ac|o.
a
•<|ti.|co
94
I
5 §
T3 M
^H O
-H O
3 V1
e ^
O b
so <r O r-* r^. r* -a-
H ■* H CM iH H O
>» o o ^o r^ vo r*
-j m m o o o
r-» oo o co oo -J O
(■n iH m CM O
co n O co o O o>
(N H * O H ' O
cr* oo en m
.h m o> en
c^ h co rl
*$■
o
CO
o
m
CM
o*\
m
H
CM
00
r*
Ln
m
rj
rs
v£>
O
^H
m.
<r
O
m
m
m
CNJ
H
O
vO
9i
•H
m
O
CO
H
<r
01
tg
c
M
CO
T-t
CT1
to
3
u
tO
9
m
60
0
m
-J
C
o
R
u
Xi
AJ
to
B
9
<ll
X
t-.
c
o
a
tH
JJ
•H
3
CO
CD
x
eg
•o
u
X
-n
to
u
■o
u
c
AJ
4-J
3
a
in
to
U
ttl
to
o
Q.
to
>-,
m
u
B
>,
to
to
x:
0.
u
>H
m
• • -C
>
<
o
U
<
Cu
OO
Efl
IT
9
3
C
u
M
o
u
4-1
X
o
X
01
u
M
u
B
U
to
to
to
>.
9
9
x:
§
o
o
B
, • u
1
oo
CO
-)
<r -J <r ao r-»
o m* d d d
4J
m
aj
s
(J
tn
9
AJ
u
u
C
M
O
Hi
o
4-1
B
T3
4J
o
3
-^
u
11
U
01
♦J
■o
aj
4J
<l
si
•I
-■ °
pq|cu
O i— ( f— I CN CN sD
O m <H -tf rH
O -d- m \£> en
o o o -3- oo m
m oo en m
m o> in in m
O en .-( o r-*
ON HNO 00
r* .-i cm m en
O-^COONO
oo oo -^- <r
O in en i— i r—
o o O m o O
O ** rs \D M N
en «h o in o
<h o m .-h
O O CT. O vD
<■ ^ N H CO
O vO H iN n
O O CO NN O
m o O O O o
>-t r- m i-h
n o> h
d *n oo
o> on co o o o
d d o>
0-J»OH
CM »tf CM CT»
o o oo o o m
•^ \£> r-» CM
pkKIu
j
00
a
k-i
3
til
aj
XI
eg
T)
U
C
tg
a
to
to
to
. o
<
PL.
cc fco wt <|<|ui|ou.|lo| h u|<|<|co|ci4| u|to|cu|w|pa
m o cm -* en
O CO H COOMA
m »H rH vj m
en cm oo o in i
O r-* -a* ^O in
ITI rH
CO
CO
CO
00 \D
lC
<r
o
n lO
C4
•a-
SO
r^ oo
CO
CO
o
tM t>\
ON
vO
-» 00
m
1
1
eg
to
o
tn on
m
■*
m
o rsi
H
pH
o
R
u
3
U
u
O
i~j
a
u
3
0)
*-J
■o
POlU
tn
4-J
<T3
iJ
d
u
cu
o
T7
J->
-^-1
u
Ui
•
4-J
u
<
BQl
95
Table 2 — Number and percent sprouting following crown removal by fire of important shrub species in Lava
Reds National Monument
Date
sampled Unit name
Date
Species
No . No . No .
shrubs shrubs shrubs Percent
preburn burned sprouted sprout
7-8-76 Upper Ice Cave I
6-27-75 Purshia tridentata 308 308 60 19
Haplopappus bloomeri 57 57 57 100
Ribes cereum 1 11 100
8-16-76 Upper Ice Cave II
8-22-75
P. tridentata
H. bloomeri
208
22
208
22
2
21
1
95
9-11-76 Upper Ice Cave III 11-3-75
P. tridentata
H. bloomeri
_R. cereum
Artemisia tridentata
512
28
5
10
407
24
5
0
85
21
5
21
88
100
6-26-79 Caldwell Butte
6-28-78 Headquarters
6-27-79 Cillem's Camp II
5-28-77 P. tridentata 855 735
Ribes velutinum 40 40
R. cereum 18 16
H. bloomeri 17 16
Chrysothamnus nauseosus 11 9
' 7-8-77 A. tridentata 333 246
P. tridentata 85 61
Cercocarpus ledif olius 17 7
JR. velutinum 2 1
JR. cereum 2 1
Chrysothamnus viscidiflorus 3 2
C. nauseosus 4 4
Prunus emarginata 2 2
7-6-77 C. nauseosus 197 128
6 4
C. viscidiflorus
43
22
9
14
3
0
0
0
0
0
0
0
2
1/20
6
55
56
88
33
0
0
0
0
0
0
0
100
45
100
6-27-79 Sheep Enclosure
7-5-78
nauseosus
Tetradymia canescens
R. velutinum
44
8
2
— Only crowns with live foliage above 20 cm in height were included in preburn inventory — sprouts of
small plants exceeded 20 cm in postburn inventory.
Crown mortality of bitterbrush was nearly complete
except on the 1977 burn where 11 percent survived.
Shrub and grass cover represented between 4 and 25
percent preburn cover values within the first 2 to
3 years.
Rabbitbrush goldenweed, Haplopappus bloomeri,
responded vigorously to fire. Within 3 years, its
percent cover on the units burned on August 27 and
November 2, 1975, exceeded preburn values.
Eighty to 100 percent of the bitterbrush crowns
within 24 macroplots were killed by fire, but as
much as 20 percent of these resprouted from root
crowns by the following year (table 2). Shrubs
burned in early summer and fall responded with
higher sprouting rates.
Of the bitterbrush plants that were top killed and
sprouted from root crowns, many had flowered and
produced seed in the first 2 to 3 years. On the
Caldwell Butte unit burned in May 1977, 65 percent
of the shrubs that sprouted had produced seed by
the summer of 1979. All of the 50 established
bitterbrush plants examined adjacent to the burn
units produced abundant seeds. Several species
occurring on 0.03-ha macroplots are not known to
sprout from adventitious buds or ligno-tubers,
such as mountainmahogany and western juniper.
Mountainmahogany does have the ability to sprout
from aboveground axillary buds, however, if roots
and cambium survive. Nearly all of the rabbitbrush
goldenweed and wax currant, Ribes cereum, plants
sprouted following crown kill on the units burned
in 1975. Rabbitbrush goldenweed, wax currant, and
desert gooseberry, Ribes velutinum, sprouted 88,
63, and 46 percent, respectively, the year
following the 1977 Caldwell Butte fire and 88, 56,
and 55 percent by two summers following burning.
96
Table 3 — Mortality of ponderosa pine from prescribed burn on 28 May 1977, Lava Beds National Monument
Stems per
hectare
Diameter breast high
-centiraeters-
Caldwell Butte (28 May 1977)
Preburn live
Postburn live
Trees killed
Upper Ice Cave (27 June 1975)
Preburn live
Postburn live
Trees killed
Upper Ice Cave (22 August 1975)
Preburn live
Postburn live
Trees killed
387
338
49
142
126
16
69
49
20
21.8
23.6
13.7
32.9
36.2
6.0
56.6
69.2
24.5
7.80
8.55
7.29
29.90
30.10
3.10
31.70
23.50
27.20
Upper Ice Cave (3 November 1975)
Preburn live
Postburn live
Trees killed
66
62
4
34.4
36.0
4.9
24.2
23.8
2.04
mean.
standard deviation.
Overstory ponderosa pine on all plots was not
greatly affected by understory burning (table 3);
generally only the smaller diameter trees were
killed. Between 6 and 29 of the trees were killed
by the four fires. The average d.b.h. of those
trees killed was significantly less than the
preburn stand average ( <*<_ .01) following all four
burns (table 3).
Mountainmahogany-Big Sagebrush Association
Only one prescribed burn was conducted in this
association where transects and macroplots were
burned. This was the Headquarters unit which was
burned on July 7 and 8, 1977.
Plant crowns were killed by fire along 16 of the 25
permanent transects established prior to burning.
The other nine were used as control transects.
Total perennial crown cover along the permanent
transects was 35 percent before the fire (table 1).
Following the fire, crown cover was reduced to 16.1
percent or 46 percent of preburn cover. Bunchgrass
cover was reduced from 5.9 to 5.1 percent while
shrub and tree cover was reduced from 29.1 to 11.0
percent. Mountainmahogany cover was reduced 64
percent by the fire. Western juniper cover was
reduced 82 percent.
Within seven 0.03-ha macroplots, 126 of 448 per-
ennial shrubs (28.6 percent) survived the burn
(table 2). Two of the surviving shrubs were bitter
cherry, Prunus emarginata, that had sprouted from
root crowns. They were the only shrubs to sprout.
Although bunchgrasses appeared to have a lower
mortality rate than shrubs (36.2 versus 22.1
percent), they showed poor vigor in the 1978
season and responded poorly to burning.
Shrub/Grass Association
Within the shrub/grass associations, four pre-
scribed burns were conducted in 197 5 and one in
1977. The Lava Overlook units (burned June 25,
August 22, and November 2, 1975) were dominated by
big sagebrush and gray rabbitbrush with bluebench
wheatgrass, Sandberg's bluegrass, Poa sandbergii,
and needlegrasses as principal subordinates before
burning (table 1). Big sagebrush was completely
killed along these transects and within the first
3 years showed no significant reinvasion. Gray
rabbitbrush crowns also were nearly completely
killed by these burns but showed some evidence of
regrowth along burned transects.
The bunchgrass crowns on the units were consumed
by the fires, but plants responded vigorously
following burning compared to those on the control
transects. Crown cover of bunchgrasses on the
control plots declined during the 3 years follow-
ing burning, while bunchgrass cover on the unit
burned in November doubled by the 2d year and had
greater cover than preburn. The bunchgrasses on
the unit burned in August showed the least vigor
following burning; crown cover was only 87 percent
of preburn value by the 3d year. '
97
Bluebunch wheatgrass and Thurber's needlegrass
responded more vigorously than other bunchgrasses.
On units burned in June, July, and November, cover
exceeded preburn levels within the first 3 years
following burning and showed significant increases
compared to the control transects.
Gray rabbitbrush crowns in the Gillem's Camp II
unit burned on July 6, 1977, were nearly completly
killed by the fire (table 1). They showed little
recovery within the first few years.
Bottlebrush squirreltail occurred along three
transects on the Gillem's Camp II unit that were
burned during July 6. By 1978, crown cover had
regained 61 percent of preburn level; but along
four unburned control transects, cover was 120
percent of the preburn level.
The percent cover of cheatgrass doubled from
preburn to 1978 on the five transects in the burn
units but was 430 percent of the preburn level
along three control transects.
Only 35 percent of the shrubs within two 0.03-ha
quadrats survived the July 7, 1977, burn (table 2).
Later that summer, only two green rabbitbrush,
Chrysothamnus viscldif lorus , shrubs had resprouted.
Since only those shrubs greater than 20 cm in
height were included in preburn inventory, small
and very decadent shrubs were not included. By
1979, the number of green rabbitbrush shrubs
greater than 20 cm within the macroplots was four
times the preburn count.
The Homestead Flow unit, burned June 26, 1975, was
dominated by bunchgrasses, principally bluebunch
wheatgrass (table 1). The crowns of the bunch-
grasses were removed by fire along four transects.
By 1976 they had regained 61 percent of preburn
cover and 67 percent by 1978. Grasses along
control transects lost 45 percent of their cover
from 1975 to 1978. Giant wildrye, needle and
thread, and bluebunch wheatgrass all regained
nearly 60 percent of preburn cover by the year
following the fire, while Sandberg's bluegrass
regained less than 25 percent of its preburn cover.
Cheatgrass Association
Three cheatgrass-dominated lowland units on the
north end of the Monument were burned from 1975 to
1978. Two of these units were at least partially
reburned the following year. All units were
burned in June when seed was mature and beginning
to drop. None of the units showed a significant
change in cheatgrass cover compared to cover on
the control area (table 1); but the number of
viable seeds per square meter on the unit burned
on June 26, 1975, was significantly reduced
(<r£ .05) compared to control plots (table 4).
Other annuals such as f ilaree, Erodium cicutarium,
tumble mustard, Sisymbrium altissimum, and Draba,
Draba verna, also showed no significant change in
cover values.
DISCUSSION
Results from this study indicate that perennial
bunchgrasses and annuals within the Monument
respond vigorously following prescribed fire. On
the other hand, perennial shrubs such as big sage-
brush and mountainmahogany recovered very slowly.
Shrubs such as bitterbursh and gray rabbitbrush
had the ability to sprout from latent buds at the
root collar, but the amount of sprouting depended
on phenologic and climatic conditions both before
and after the fires. Other shrubs in the Monument
such as green rabbitbrush, elderberry, Sambucus
cerulea, bitter cherry and chokecherry , Prunus
virginiana, had nearly 100 percent sprouting
following fire.
Mountainmahogany has a thin bark and its cambium
is easily killed by fire. Western juniper and
ponderosa pine, on the other hand, have relatively
thick bark and are much harder to kill by fire.
Ponderosa pine and western juniper have very flam-
mable foliage; but by the time ponderosa pines
reach pole-size, their crowns are high enough that
they are quite resistant to surface fires, if
fuels are not allowed to build up under them.
Much of the literature concerning response of
vegetation to fire has been contradictory, if not
misleading. This is particularly true in the case
of bitterbrush. Hormay (1943) reported that
bitterbrush areas in California destroyed by fire
were taken over by cheatgrass and made no mention
of bitterbrush1 s ability to sprout. Blaisdell
(1950), however, reported that on the upper Snake
River plains 49 percent of the bitterbrush plants
sprouted on an area lightly burned, 43 percent in
an area moderately burned, and 19 percent in an
area heavily burned. Woolfolk (1958) reported
that only one location in California is known
where more than 25 percent of the bitterbrush
plants sprouted, while Driscoll (1963) found that
as many as 80 percent of the bitterbrush shrubs
sprouted following burning in central Oregon
gravelly loam soils. Soil moisture probably plays
the most important role in the ability of bitter-
brush to sprout following fire (Nord 1965, Britton
1979). The results of this study also suggest
that soil moisture during burning determines
bitterbrush survival. Therefore, spring and fall
fires when soil and litter are relatively moist
favor bitterbrush sprouting compared to mid-summer,
when sprouting is near zero.
Soil heat transfer characteristics, phenological
stage, and subspecies characteristics, however,
are probably also important influences on the
sprouting ability of bitterbrush following burning.
Zero to 45 percent of the gray rabbitbrush shrubs
sprouted following crown removal. Additional
information will be necessary to establish what
determines this species' ability to sprout, since
the units it occurred on were all burned at the
same time of year.
98
Table 4 — Viable seeds on burned and unburned cheatgrass (Bromus tectorum) plotsi'
Species
Number
of
samples
Pre burn
Postburn
Control
West Wildlife
Overlook Burn (6-25-75):
Bromus tectorum
Erodium cicutarium
Gillem's Camp II (7-6-77)
Bromus tectorum*
11
Number of viable seeds
per 1/10 square meter
9
486
77.4
337
425.4
529.0
380.0
9
136
73.3
135
80.5
90.4
37.4
1,149
837.2 2,516.0 6,198.2
_'x = mean; s = standard deviation; * = significant difference (cc<0.05).
Past research concerning the effects of fire on
perennial bunchgrasses indicated that phenological
stage at the time of burning and fire intensity
were of considerable importance in determining
post-fire vigor (Daubenmire 1968). The effect of
clipping of bluebunch wheatgrass was negligible
during early spring and late fall in Dubois, Idaho
(1 670-m elevation). Reduction in production
followed late May to early June clipping
(Blaisdell and Pechanec 1949). In northeast
Oregon, bluebunch wheatgrass showed little
difference in vigor following burning at various
intensities while Idaho fescue showed considerable
reduction in survival with increasing intensity
(Conrad and Poulton 1966). Both bluebunch
wheatgrass and Thurber's needlegrass had
significantly reduced leaf length the spring
following an August wildfire in south-central
Washington, but bluebunch wheatgrass had longer
culms and spikes compared to unburned plants
(Uresk et al. 1976). In this study, bunchgrasses
showed the greatest increase in cover following a
November fire. Bunchgrasses on the unit burned in
August under hot, dry conditions showed a decrease
in cover. Bluebunch wheatgrass in particular
showed a dramatic increase in cover following fall
burning. On the other hand, late June burning
reduced cover the following year by half; but
bunchgrasses regained 93 percent of their preburn
cover by the 3d year following burning. Three
years after an August burn, bluebunch wheatgrass
cover was 55 percent of the preburn value. Evi-
dently, the June burn reduced vigor but didn't
kill the crowns; while August burning killed the
crowns and lowered vigor.
On all three Lava Overlook burns, percent cover of
Thurber's needlegrass had exceeded preburn levels
by the 3d year following the fire. By 1978, cover
of Sandberg's bluegrass was less than 60 percent
of preburn values on units burned in late June,
early July, and August; while on the control
units, it increased during the same period. By
the 2d year following the November fire, its cover
exceeded the preburn value.
Cheatgrass is well adapted to fire (Klemmedson and
Smith 1964, Young et al. 1976). Though the number
of viable seeds and number of plants was reduced
by burning in June just as seed began to drop,
vigor and thus percent cover was not reduced.
CONCLUSIONS
Time of year, moisture regime, and fire intensity
all must be considered when developing fire pre-
scriptions to achieve desired results.
Bunchgrasses respond best to late summer and fall
burning when heat penetration into the crown is
low and nutrients are stored in the roots. Methods
other than just spring burning are necessary to
control cheatgrass.
Shrubs such as bitterbrush and gray rabbit brush
are more sensitive to soil moisture and fire
intensity than phenological condition when
compared to bunchgrasses.
Both spring and fall burning can result in
relatively high sprouting rates. Green rabbit-
brush, rabbitbrush goldenweed, Ribes sp. , and
bitter cherry readily sprout following fire. Fire
intensity and season of burn seem to affect bitter
cherry and rabbitbrush goldenweed very little.
Big sagebrush is easily killed by fire.
It is difficult to replicate conditions of a fire
and individual fires are often highly variable.
Additional information is necessary so managers
can develop prescriptions favorable for desirable
native species.
99
LITERATURE CITED
Blaisdell, James P.
1950. Effects of controlled burning on
bitterbrush on the upper Snake River plains.
USDA For. Serv. Res. Pap. INT-20, 3 p. Intermt.
For. and Range Exp. Stn. , Ogden, Utah.
Blaisdell, James P., and Joseph F. Pechanec.
1949. Effects of herbage removal at various
dates on vigor of bluebunch wheatgrass and
arrowleaf balsam root. Ecol. 30(3) :298-305.
Britton, Carlton M.
1979. Fire on the range.
5(4):32-33.
Western Wildlands
Canfield, R. H.
1941. Application of the line interception
method in sampling range vegetation. J. For.
39:388-394.
Conrad, C. Eugene, and C. E. Poulton.
1966. Effect of wildfire on Idaho fescue and
bluebunch wheatgrass. J. Range Manage.
19(3):138-141.
Daubenmire, Rexford.
1968. Ecology of fire in grasslands. Adv.
Ecol. Res. 5:209-266.
in
Driscoll, Richard S.
1963. Sprouting bitterbrush in central Oregon.
Ecol. 44(4):820-821.
Hormay, A. L.
1943. Bitterbrush in California. USDA For.
Serv. Res. Note 34, 13 p. Calif. For. and Range
Exp. Stn.
Johnston, A.
1957. A comparison of the line interception,
vertical point quadrat, and loop methods as
needed in measuring basal area of grassland
vegetation. Can. J. Plant Sci. 37(l):34-42.
Klemmedson, James D. , and Justin G. Smith.
1964. Cheatgrass (Bromus tectorum L. ) .
Rev. 30(2)226-262.
Bot .
Nord, Eamor C.
1965. Autecology of bitterbrush in California.
Ecol. Monog. 35(3):307-334.
Uresk, D. W. , J. F. Cline, and W. H. Rickard.
1976. Impact of wildfire on three perennial
grasses in south-central Washington. J. Range
Manage. 29(4) :309-310.
Woolfolk, E. J.
1958. Project WSTR. Crane Range restoration
segment. Western Assoc, of State and Game Coram.
Western Browse Res. 4(l):l-9.
Young, James A., Raymond A. Evans, and
Ronald A. Weaver.
1976. Estimating potential downy brome
competition after wildfire. J. Range Manage.
29(4):322-325.
100
Single- Year Response of Breeding Bird Populations to Fire in a Curlleaf
Mountainmahogany-Big Sagebrush Community
Tamara E. Tiagwad, Craig M. Olson, and Robert E. Martin
ABSTRACT
Breeding bird territories were mapped on four 16.2-hectare (ha) grids in curlleaf mountainmahogany , Cerco-
carpus ledif olius,-big sagebrush, Artemisia tridentata, communities in Lava Beds National Monument, Cali-
fornia, in the spring and summer of 1979. Grids were located in a 1973 wildfire, a 1977 prescribed burn,
and control areas adjacent to each burn. Prior to burning, all areas had scattered ponderosa pine, Pinus
ponderosa, and western juniper, Juniperus occidentalis , but mountainmahogany, antelope bitterbrush,
Purshia tridentata, and big sagebrush dominated the vegetation.
Only one small island of original vegetation survived on the wildfire grid; the rest of the area was
dominated by bunchgrasses. Vegetation structure was more varied on the control grid. The wildfire grid
also showed less avian diversity than its control. Ground nesting birds dominated the burn area, but the
control area had a variety of shrub and tree nesting birds.
The prescribed burn had an interspersion of burned and unburned areas; therefore, much "edge" and nesting
habitat. It had a wider range of ground, shrub, and tree nesting birds than the control.
Tamara E. Tiagwad, biological aid, Craig M. Olson
and Robert E. Martin, research foresters, U.S.
Department of Agriculture, Forest Service, Pacific
Northwest Forest and Range Experiment Station,
Silviculture Laboratory, Bend, Oregon.
101
INTRODUCTION
In Lava Beds National Monument, California,
wildfires historically played a major role in vege-
tation distribution and structure. Fire suppres-
sion, heavy livestock grazing, and a western pine
beetle, Dendroctonus brevicomis, outbreak have mod-
ified vegetation patterns in the last half century.
A prescribed burning program is currently being
developed in the Monument to restore presettlement
vegetation and reduce unnatural fuel accumulations.
Few studies have compared the response of breeding
bird populations to wildfires and prescribed burns
in similar habitat. The objective of this study
was to document avian density, species
composition, and avian diversity in wildfire,
prescribed burned, and unburned areas dominated by
mountainmahogany , big sagebrush, and bitterbrush.
THE STUDY AREA
The study area was located in the Lava Beds Nation-
al Monument, about A8 kilometers (km) southwest of
Tulelake, California. The bird census grids were
located in T. 45 N. , R. 4 E. , Sections 27, 28, and
29, at an elevation of about 1 454 meters (m) with
east-northeast aspects. Soils are derived from
pumice and basalt. These relatively young soils
are characterized by a stone pavement overlying
gravelly sand to sandy loam with little horizon
development. Soils are shallow with numerous
basalt outcrops.
The Monument has a modified maritime climate with
warm dry summers and cool wet winters. Weather
records from the Indian Wells Weather Station
located between study areas indicate an average
annual precipittion of 34 centimeters (cm), July
daily mean high temperatures of 27 °C, and January
daily mean low temperatures of -6°C.
Four 16.2-ha census grids were located in curlleaf
mountainmahogany-big sagebrush communities. One
of the grids was in an area burned by a 205-ha
wildfire on July 11, 1973, and another was located
in a 120-ha area prescribed burned on July 7 and
8, 1977. From aerial photographs and ground
surveys, unburned control grids were selected
approximately 0.2 km from each burned unit with
plant communities similar to those characterizing
the burned grids prior to the fires. Until the
1973 wildfire, no fire over one-quarter acre had
occurred in the study area since the 1930' s.
Although the wildfire and prescribed fire occurred
at the same time of year, had similar daytime high
temperatures (28° and 27°C), and had similar low
humidities (20 and 18 percent, respectively), they
differed significantly in winds and fuel moisture
content during the fire. The wildfire followed a
month with no rain and low humidity while the
prescribed fire followed a month of exceptionally
high rainfall (5.1 cm). Because of this, live
fuels during the prescribed fire probably had much
higher moisture content. Also, dead wood between
0.62- and 2.54-cra diameter had a moisture content
of 3 percent during the wildfire; but during the
prescribed fire, it had a moisture content of 6
percent. Winds associated with the dry lightning
storm that started tha 1973 wildfire were gusty
and were probably around 15 to 20 kilometers per
hour (km/hr) compared to winds of 0 to 12 km/hr
during the prescribed fire. The combined effect
of relatively high winds and low fuel moisture
content significantly influenced fire behavior and
subsequent vegetation patterns.
METHODS
Vegetation
Vegetation was sampled at 30 systematically locat-
ed grid points on each 16.2-ha bird census grid.
Alternate grid points were sampled on alternate
grid lines. A rope 28.35 m long was stretched
from a grid point in a randomly chosen compass
direction and vegetation was sampled along the
transect by two methods. First, ground cover of
shrubs and trees was measured by the line inter-
cept method (Canfield 1941). Length of the line
intercept, average height, and an estimate of
percent of dead crown were recorded for each shrub
and tree. Dead plants, both standing and down,
were not included in the sampling. Second,
frequency of grasses, shrubs, and trees was deter-
mined while measuring ground surface cover as
described by Daubenmire (1968). Since vegetation
was sampled following the breeding season, annual
forbs were beyond recognition and not included in
the analysis. Fifteen subplots on each transect
were sampled at 1.9-m intervals along the rope
with a 20- x 50-cm steel frame placed at right
angles along one side of the rope. Coverage was
visually estimated for bare soil, plant litter,
stones over 2-cm diameter, logs over 7.6-cm
diameter, and live plants.
Constancy and frequency were calculated for all
grasses, shrubs, and trees. Constancy is the
percent of transects along which a species occurs.
Frequency is the percent of quadrants within which
a species occurs.
Bird Census
Territorial male birds were censused during the
breeding season by the spot-mapping method
(International Bird Census Committee 1970).
Censusing began April 30, 1979, and ended July 19,
1979. All of the censuses were done between 0445
and 0900 Pacific Daylight Time and were 1 to 3
hours long. Temperature, wind speed, and cloud
cover were recorded before and after each census.
Sixty individual censuses were made — 15 each
on the wildfire and the prescribed burn control
grids, 16 on the wildfire control grid, and 14 on
the prescribed burn grid. Each grid was censused
one or two times per week.
The square grids cover 16.2 ha, grid lines were
spaced 36.6 m apart, and grid points were marked
by a permanent iron rod with a letter and number
corresponding to its relative position on the
grid. Three of the grids had 144 (12 x 12) grid
points and the wildfire control grid had 169
(13 x 13) grid points with grid lines spaced
33.5 m apart. The narrower spacing was to allow
for lower bird visibility in that grid.
102
The censusing technique was to walk slowly along
the grid lines, alternating starting points to
avoid favoring a particular part of the grid, and
identify and record the species, positions, and
movements of all- birds on a map of the grid. The
daily maps were compiled on individual species
maps covering the entire census period for each
grid, and clumped observations were used to
delineate territories. At least two separate
observations of a territorial nature on separate
days, such as singing or aggression toward other
birds, and one of a general nature such as
foraging were used as criteria for determining
territories. Birds with territories only partly
within a grid area were included in density and
diversity calculations which would slightly
inflate these values. The number of territories
per 16.2 ha, was converted to territories per 40.5
ha (100 acres), a commonly used figure in bird
density studies (Gashwiler 1977, Feist 1968).
Diversity values were calculatd from th~ original,
unconverted data using two formulas:
s
(1)
H'
P. log P. (Shannon 1948)
'—. — r i e i
i=l
where P^ is the proportion of species i of the
total number of individuals in the populations,
s is the number of species, and H' is the dimen-
sionless information index and:
_s_ 2 /
I = N -> n /N -^N (Mcintosh 1967) (2)
i=l i/
where N is the total sample size, s is the number
of species 1 represented in the sample, n is the
number of indivduals of a particular species in
the sample, and Ip is the percent of theoretical
maximum diversity for a particular N. In both
formulas, the larger the index value, the greater
the diversity. Both indices of diversity were
calculated, because they provide different
information.
Cover of shrubs was 5.3 percent, with bitter
cherry, Prunus emarginata, accounting for 73
percent of the total shrub cover.
A small area of vegatation, largely mountain-
mahogany and western juniper occurring on 30 and
33.3 percent of the transects with average heights
of 281 cm and 532 cm, respectively. The control
grid had a total shrub coverage of 23.3 percent,
with big sagebrush (10.0 percent) and bitterbrush
(8.9 percent) forming 81 percent of the total.
Bluebunch wheatgrass, cheatgrass, and Idaho
fescue, Festuca idahoensis, were the most abundant
of the seven grass and sedge species with
constancy values of 100, 96.7, and 73.3 percent,
respectively. Cheatgrass had the highest
frequency value (45.8), which was half that of the
wildfire grid.
1977 Prescribed Burn Grid
The prescribed burn covered only 37 percent of the
sampling grid, resulting in much edge and a high
percent of tree and shrub cover. Mountainmahogany
had a coverage of 19.3 percent and occurred on
73.3 percent of the transects. Western juniper
had a coverage of 5.2 percent and a constancy of
46.7 percent. The total low shrub cover was 6.5
percent with big sagebrush, the dominant shrub,
covering 4.3 percent and antelope bitterbrush 21.1
percent. They comprised 66 and 32 percent of the
total low shrub cover, respectively. Of the eight
grass and sedge species occurring on the prescribed
burn, Idaho fescue and bluebunch wheatgrass were
the most abundant with frequencies of 42.9 and
40.2 percent, respectively. These two species
along with cheatgrass, Thurber's needlegrass,
Stipa thurberiana, and Sandberg's bluegrass, Poa
sandbergii , occurred along more than half the
transects.
RESULTS
1977 Prescribed Burn Control Grid
Vegetation
Major differences in the estimates of cover of
trees, shrubs, and grasses occurred among burned
and control areas (tables 1, 2, and 3). Bunch-
grasses dominated the wildfire grid except for a
small island of shrubs and trees that survived the
fire. The wildfire had a very high percent of
rock cover and low soil cover compared to the
other three areas. Dominant species on the
wildfire control grid were mountainmahogany,
antelope bitterbrush, and big sagebrush. The
prescribed burn grid has an intersperion of burned
and unburned areas. Dominant species for this
grid and its control were the same as those found
on the wildfire control grid, and all four areas
had similar cover of litter.
1973 Wildfire Grid
Five percent of the total area within the wildfire
grid did not burn. The grid was dominated by
three grasses, constancy value for each were:
cheatgrass, Bromus tectorum, 100 percent;
bluebunch wheatgrass, Agropyron spicatura, 80
percent; needlegrass, Stipa spp. , 46.7 percent.
Cheatgrass had the highest frequency value (94.5).
Tree and tall shrub cover on the control grid was
32.5 percent, mountainmahogany covered 24.6
percent, and western juniper covered 7.9 percent
of the area. The prescribed fire control grid had
a tree and tall shrub cover 2.3 times that of the
wildfire control grid.
Total low shrub cover on the control grid was 15.4
percent. Big sagebrush (9.2 percent) and antelope
bitterbrush (6.2 percent) were the only shrubs
present on the transects. The total shrub cover-
age on the 1977 prescribed burn control grid was
66 percent of that on the 1973 control grid.
Grasses and sedges with constancy values above 50
percent were Idaho fescue (100 percent), bluebunch
wheatgrass (96.7 percent), Thurber's needlegrass
(96.7 percent), cheatgrass (90 percent), and
Sandberg's bluegrass (65.7 percent). Idaho fescue
had the highest frequency (45.6 percent) of the
nine species occurring along the grid transects.
103
Bird Census
Bird census data indicated some large differences
in bird populations (table 4, 5, and 6). The 1973
wildfire grid had the lowest number of bird terri-
tories (41/40.5 ha), the lowest number of territor-
ial species (6) (table 3), and the lowest diversity
values (table 5). Total number of species occur-
ring on the grid was 25. Western meadowlarks,
Sturnella neglecta, rock wrens, Salpinctes
obsoletus, scrub jays, Aphelocoma coerulescens,
and mourning doves, Zenaida macroura, comprised
24, 24, 24, and 12 percent, respectively, of the
total breeding population on the grid. Ground
nesters made up 49 percent of the total breeding
population (table 4).
The wildfire control grid had the highest number
of bird territories (70/40.5 ha), the highest
number of territorial species (14), and the highest
total number of species (27) of the four areas.
Tree nesters and shrub/tree nesters comprised 39
percent and 34 percent of the total breeding
population. Scrub jays had the highest breeding
bird density (13 pairs/40.5 ha) on the grid. This
grid also had the highest diversity values (H' =
.86 and Ip = 2.53) and the highest average number
of species observed per census (9.9).
The number of breeding bird territories on the
1977 prescribed burn grid was 61 per 40.5 ha.
Total number of species occurring on the grid was
25, with 11 species defending territories.
Shrub/tree nesters made up 43 percent of the total
breeding population, and tree nesters formed 36
percent of the total. Scrub jays, common
bushtits, Psalt riparus minimus, and American
robins, Turd.us migratorius, were the most abundant
species with 13, 10, and 8 pairs per 40.5 ha.
The prescribed burn grid had diversity values of
H' = .81 and Ip = 2.23.
The estimated number of breeding bird territories
on the 1977 control grid was 49 per 40.5 ha. This
control grid had the lowest total number of
species occurring on the grids (23), 8 of which
established territories. Shrub/tree nesting
species were 57 percent of the total breeding
population, and tree nesters were 22 percent of
the total. More than half of the breeding birds
were scrub jays or chipping sparrows, Spizella
passerlna. This grid had the lowest average
number of species observed per census (6) and had
diversity values of H' = .77 and Ip = 1.94.
DISCUSSION
Differences in breeding bird density, species com-
position, and diversity values between burn and
control plots suggest that wildfires and prescribed
burns affect habitat suitability for many avian
species. Only one species, the scrub jay, nested
on all four grids. Three shrub and tree nesting
species, the common flicker, Colaptes avratus,
common bushtit, and blue-gray gnatcatcher, Poli-
optila caerulea , nested on all but the 1973
wildfire grid.
Though the two control grids had similar vegeta-
tion, the vegetation analysis indicates several
differences in frequency or distribution.
The 1973 control had higher percentage cover of
low shrubs than did the 1977 control (23 versus
15.4 percent) and lower percentage cover values
for mahogany and western juniper (14 versus 32.5
percent). Consequently, the wildfire and pre-
scribed burn grids were not directly compared.
Breeding bird density and diversity values were
considerably higher on the 1973 wildfire control
grid than on the wildfire grid. The control grid
also had much higher estimates for total shrub
coverage and percent cover of the two most abundant
shrub species. This correlates with the substan-
tial number of shrub nesting species on the grid.
Gashwiler (1977) reported an average of 91 breeding
bird territories per 40.5 ha in an area in central
Oregon and shrub species and coverage similar to
the wildfire control grid. In his study, average
numbers of territorial species were 6.3, with
Brewer's sparrows, Spizella breweri, sage sparrows,
Amphispiza belli , horned larks, Eremophila alpes-
tris, and sage thrashers, Oreoscoptes montanus,
the most abundant. The wildfire control grid had
70 territories per 40.5 ha, 23 percent fewer than
the average density Gashwiler (1977) found. The
number of territorial species on the control grid
was over twice the average number reported in the
central Oregon study. Species composition was much
different for the two areas. Scrub jays, common
bushtits, and sparrows were most abundant on the
control grid. These difference in estimates of
breeding bird density could be the result of
vegetative differences, seasonal fluctuation, or
sampling error.
The wildfire grid had comparatively low bird
density and avian diversity values. Most of the
breeding birds were ground nesters. The low
diversity of vegetation on the grid limited the
range of niches available for nesting birds. Avian
species diversity has been positively correlated
with foliage height diversity by MacArthur and
MacArthur (1961). Since foliage height diversity
tends to increase as succession increases; avian
species diversity is usually low in grassland
habitats.
No trees were encountered along wildfire grid vege-
tation transects, yet we had two tree nesting birds
breeding on the grid. One of these, a mountain
bluebird, Sialia currucoides , probably used a
cavity in one of the standing dead western junipers
not included in our analysis. The other, a robin,
nested in the patch of shrubs and western junipers
spared by the fire. The dead western junipers
throughout the area also provided excellent singing
and roosting perches for western meadowlarks,
Sturnella neglecta, mourning doves, and scrub jays.
104
Table 1 — Cover percent, plant number, height, and percent of dead crown for the 1973 wildfire grid and its
control
1973 Wildfire 1973 Control
Percent Percent Percent Percent
cover Height (cm) dead cover Height (cm) dead
Number Number
of y 2/ of
Species samples x s x s 5J s samples
Ground cover
Rock
—
Soil
—
Litter
—
Log
—
Live plant
Low shrubs
Big sagebrush
—
Antelope bitterbrush
2
Desert gooseberry
4
Squaw currant
1
Green rabbitbrush
—
Gray rabbitbrush
4
22.0 19.9 — — — — — 4.3 7.1
9.5 13.3 — — — — — 37.2 16.2
61.0 17.2 — — — — — 55.3 16.3
0.6 0.5
9.9 3.0 — — — — — 3.4 2.3
0.2
1.1
0.2
0.6
0.1
0.3
144
77.5
31.8
35.5
48.8
67
81.3
19.3
19.3
16.8
4
45.0
—
5.0
—
2
11
73.8
16.5
3.5
1.7
17
.0.0
12.4
57.7
19.2
29.2
26.2
8.9
9.2
127.9
53.1
23.4
23.0
0.4
1.1
118.8
26.6
55.0
38.5
1.7
5.7
70.0
28.3
37.5
31.8
0.5
1.8
40.5
17.2
34.3
27.0
1.3
3.0
58.8
28.6
67.9
19.4
0.4
1.7
91.7
22.1
9.5
15.4
0.3 1.1
Bitter cherry 43 3.8 7.2 103.7 35.3 4.3 4.4
Choke cherry 3 0.5 1.6 81.7 27.5 13.3 2.9
Serviceberry 1 0.1 0.3 75.0 — 1.0 — 1 0.1 0.3 50.0 — 5.0 —
Total low shrubs 58 5.2 252 23.3
Tall shrubs
Mountainmahogany — — — — — — — 31 6.2 10.4 281.1 34.6 24.5 14.7
Trees
Western juniper — — — 16 6.5 13.5 531.6 300.1 5.1 6.3
Ponderosa pine —
Total trees
Total trees
and shrubs 58 5.2 300 37.3
1/ -
— x = mean.
2/
— s = standard deviation.
6
6.5
13.5
531.6
300.1
5.1
1
1.3
5.0
750.0
—
10.0
7
7.8
105
Table 2--Cover percent, plant number, height, and percent of dead crown for the 1977 prescribed burn grid
and its control
Species
1977 prescrlh
ed burn
1977 control
Number
Percent
cover
Height (cm)
Percent
dead
Number
of
samples
Percent
cover Height (cm)
Percent
dead
of
samples
1/ 2/
x — s —
x s
x s
x s x s
"x s
Ground cover
Rock
—
Soil
—
Litter
—
Log
—
Live plant
—
Low shrubs
Big sagebrush
72
Antelope bitterbrush
26
Desert gooseberry
2
Total low shrubs
118
Tall shrubs
Mountainmahogany
97
Trees
5.0 6.5 — — — — — 6.3 6.1
28.3 15.1 — — — — — 28.4 13.9
61.7 17.4 — — — — — 59.7 13.1
5.5 3.5 — — — — — 5.9 3.2
4.3 4.9 51.6 20.2 31.6 26.0 118
2.1 4.2 78.5 26.9 30.2 20.8 68
0.1 0.8 75.0 21.2 89.5 13.4
6.5 186 15.4
9.2
9.7
60.2
18.2
36.2
24.0
6.2
6.2
82.9
24.7
42.8
20.7
19.3 21.3 254.5 51.3 21.9 15.1 119 24.6 20.3 265.1 63.2 27.7 17.9
Western juniper 18 5.2 8.4 523.1 32.2 5.8 10.2 18 7.9 13.5 648.5 407.4 5.1 3.5
Ponderosa pine 1 0.7 3.5
Total trees 19 5.9 18 7.9
523.1
32.2
5.8
10.2
18
600.0
20.0
18
323
Total trees
and shrubs 234 31.7 323 47.9
_' s = standard deviation.
106
Table 3 — Frequency (f) and constancy (c) values for vegetation on the 4 census grids. Mean (x) and
standard deviation (s) are given for frequency
1973 wildfire
1973 control
1977 prescribed
burn
1977 control
Vegetation
c(%) 5(f) s(f) c(%) xtf) s(f) c(%) *(f) s(f) c(%) 5(f) s(f)
Grasses and sedges
Bluebunch wheatgrass
Agropyron spicatum
Rush-leaved bluegrass
Poa juncif olia
Sandberg's bluegrass
Poa sandbergii
Needlegrass
Stipa spp.
Thurber's needlegrass
Stipa thruberiana
Bottlebrush squirreltail
Sitanion hystrix
Ross' sedge
Carex rossii
Idaho fescue
Festuca idahoensis
Prairie junegrass
Koeleria cristata
Cheatgrass
Bromus tectorum
Low shrubs
Big sagebrush
Artemesia tridentata
Antelope bitterbrush
Purshia tridentata
Desert gooseberry
Ribes velutinum
Squaw currant
Ribes cereum
Green rabbitbrush
Chrysothamnus viscidiflorus
Gray rabbitbrush
Chrysothamnus nauseosus
Bitter cherry
Prunus emarginata
Choke cherry
Prunus demissa
Rabbitbrush goldenweed
Haplopappus bloomeri
Tall shrubs
Mountainmahogany
Cercocarpus ledifolius
Trees
Western juniper
Juniperus occidentalis
Ponderosa pine
Pinus ponderosa
80.0 19.4 16.8 100.0 39.8 24.0
20.0 1.6 3.4
46.7 8.5 13.5 56.7 11.2 13.2
10.0 0.9 2.9 26.7 5.2 11.9
20.0 2.0 4.3
13.3 1.8 5.3 73.3 12.5 14.8
100.0 94.5 8.6 96.7 45.8 30.7
63.3
22.9
23.8
10.0
0.9
2.9
70.0
13.8
13.8
10.0
0.7
2.1
13.3
0.9
2.4
6.7
0.5
1.8
16.7
2.9
7.7
13.3
0.9
2.4
26.7
3.1
6.3
33.3
7.1
12.4
3.3
0.4
2.4
10.0 0.7
2.1 —
93.3 40.2 25.1 96.7 39.1 17.9
3.3 0.2 1.3 20.0 2.0 4.7
73.3 17.8 16.8 56.7 6.5 7.9
70.0 13.2 15.2 96.7 39.7 25.0
16.7 1.2 2.7 36.7 2.8 3.9
13.3 2.3 5.4 30.0 4.1 7.7
80.0 42.9 30.5 100.0 45.6 23.3
43.3 6.1 7.5
66.7 18.3 24.5 90.0 35.0 23.7
63.3 14.9 15.1 70.0 19.5 19.9
33.3 4.5 9.3 63.3 12.5 12.7
3.3 0.2 1.3 6.7 0.5 1.8
3.3 0.4 2.4
30.0 7.1 12.7 73.3 22.5 21.4 80.0 29.1 24.3
33.3 8.0 15.5 46.7 6.0 7.9 33.3 7.6 12.1
3.3 0.7 3.7 3.3 0.7 3.7
107
Table 4 — Breeding bird territories per 40.5 ha for the 4 census grids
1973 1977
Species Wildfire Control Prescribed burn Control
Red-tailed hawk3 (Buteo jatnaicensis) - 3 - -
American kestrel (Falco spraverius) + + + +
California quail (Lophortyx calilornicus) +3 5 +
Mourning dove (Zenaida macroura) 5 + 5 +
Poor-will (Phalaenoptilus nuttallii) - + - -
Common nighthawk3 (Chordeiles minor) - +
Rufous hummingbird (Selasphorus rufus) + + - -
Common flicker (Colaptes auratus) +5 3 3
Western kingbird (Tyrannus verticalis) + + -
Ash-throated flycatcher (Myiarchus cinerascens) - + +
Empidonax flycatcher (Empidonax spp.) - + + -
Western wood pewee (Contopus sordidulus) + + - +
Violet-green swallow (Tachycineta thalassina) + + - -
Purple martin (Progne subis) + - -
Scrub jay (Aphelocoma coerulescens) 10 13 13 13
Black-billed magpie (Pica pica) + - +
Common raven (Corvus corax) + + + -
Mountain chickadee (Parus gambeli) - 5 + 5
Plain titmouse (Parus inornatus) - 5 5 +
Common bushtit (Psaltriparus minimus) 8 10 5
White-breasted nuthatch (Sitta carolinensis) + -
Bewick's wren (Thryomanes bewickii) - + - +
Rock wren (Salpinctes obsoletus) 10 3 - +
American robin (Turdus migratorius) 3 + 8 +
Hermit thrush (Catharus guttatus) - - - +
Mountain bluebird (Sialia currucoides) 3 + 3
Townsend's solitaire (Myadestes townsendi) + - - -
Blue-gray gnatcatcher (Polioptila caerulea) - 3 3 3
Ruby-crowned kinglet (Regulus calendula) - 3 - -
Yellow warbler (Dendroica petechia) + - -
Yellow-rumped warbler (Dendroica coronata) - + + +
Townsend's warbler (Dendroica townsendi) + - -
Wilson's warbler (Wilsonia pusilla) + + +
Western meadowlark (Sturnella neglecta) 10 - 3 +
Northern oriole (Icterus galbula) + -
Brewer's blackbird (Euphagus cyanocephalus) + + 3 +
Brown-headed cowbird (Molothrus ater) - + - -
Western tanager (Piranga ludoviciana) - - + -
Evening grosbeak (Hesperiphona vespert ina) - - - +
Purple finch (Carpodacus purpureus) +3 +
Cassin's finch (Carpodacus cassinii) + - + -
Pine siskin (Splnus pinus) + - - -
Rufous-sided towhee (Pipilo ery throphthalmus) - 5 + 5
Dark-eyed junco ( Junco hyemalis) - + 5
Chipping sparrow (Spizella passerina) - 3 +10
Brewer's sparrow (Spizella breweri) - 8 - +
Total number territories 41 70 61 49
Total number territorial species 6 14 11 8
Total number species 25 27 25 23
a = Species with large territories.
+ = The species was present on the grid but no territories were established.
108
Feist (1968) censused breeding birds on sagebrush
grassland areas in central Montana in 1966 and
1967. The vegetation type reported was roughly
similar to that on the wildfire grid. Feist found
four species with 42.5 breeding pairs per 40.5 ha
on a sagebrush range that had been sprayed with
herbicide the previous year. Dominant species were
the Brewer's sparrow, vesper sparrow, Poecetes
gramineus, and western meadowlark. The number of
species breeding on the wildfire grid was about
the same as that reported in central Montana, but
the species composition was much different. Domi-
nant species for the present study were the scrub
jay, rock wren, Salpinctes obsoletus, and western
meadowlark with six breeding species and 41 pairs
per 40.4 ha.
The prescribed burn grid had higher breeding bird
density and diversity values than its control
grid. Percent cover values for both trees and
shrubs were considerably lower on the burn grid.
The mosaic of burned and unburned areas resulted
in a greater variety of ground, shrub, and tree
nesting species on the burn grid. In central
Oregon, Gashwiler (1977) found an average of 202.7
territories per 40.5 ha over 3 years in a western
juniper type with many of the same shrub species
and similar coverage values compared to the pre-
scribed burn grid. The prescribed burn grid had
61 territories per 40.5 ha, 70 percent fewer than
Gashwiler' s (197 7) average density. The. number of
territorial species was roughly the same for both
studies. Gashwiler recorded an average of 13.7
species compared to 11 in this study. The most
abundant species on the prescribed burn grid dif-
fered from those reported in central Oregon. Gash-
wiler (1977) found mostly empidonax flycatchers,
Empidonax spp. , house finches, Carpodacus mexi-
canus, chipping sparrows, mountain bluebirds,
American robins, and mountain chickadees, Parus
gambeli . Scrub jays, common bushtits, and American
robins were most abundant on the prescribed burn
grid.
SUMMARY
This report represents only 1 year's results,
which indicate that conversion of shrubland to
grass land by the 1973 wildfire led to a decrease
in density and numbers of species of birds, a
decrease in bird diversity, and a shift in bird
species composition compared to unburned areas
with a more varied vegetation structure.
The results also suggest that heterogeneous habitat
produced by the 1977 prescribed burn led to an
increase in density and numbers of bird species,
an increase in bird diversity, and some shift in
the species composition. Thus, prescribed burning
that results in greater vegetation mosaic may
produce a greater variety of ground, shrub, and
tree nesting birds compared to unburned areas.
Table 5 — Numbers of territories per 40.5 ha and.
numbers of nesting species (in parentheses) on the
4 census grids
1973
1977
Nesting category
ground
ground/shrub
ground /shrub/ tree
shrub/tree
tree
Burn
Control
Burn
Control
20 (2)
5 (1)
10 (1)
6 (2)
6 (2)
5 (1)
8 (1)
24 (3)
27 (7)
8 (2)
5 (1)
26 (3)
22 (5)
5 (1)
5 (1)
28 (3)
11 (3)
Table 6 — Diversity values and numbers of species
observed per census for the 4 grids
1973
1977
Nesting category Burn Control Burn Control
IP
.72
.86
.81
.77
Diversity:
H'
(1.66)
(2.53)
(2.23)
(1.94)
Number of
species
x
6.4
9.9
8.5
6.0
observed
per census
s
1.6
2.2
2.5
1.1
Ip = percent of theoretical maximum diversity for
a particular sample.
H' = dimensionless information index.
x = average number of species per census.
s = number of species in a census.
Bird density and species diversity may be directly
related to age of burns (Johnson 1976). Therefore,
the effects of burning on breeding bird populations
should be documented for several seasons before
reaching conclusions about the impact of a specific
wildfire or prescribed burn. Also, dramatic sea-
sonal fluctuations of some bird populations occur
even when the vegetation is not modified (Balda
1975).
Both prescribed fire and wildfire can produce
highly variable vegetation structure depending on
the vegetation before the fire and burning condi-
tions. Because of this, both can change vegetation
diversity and structure, and thus avian diversity
and composition. Since we can pick conditions and
burning patterns for prescribed fires but not for
wildfires, intentioned use of fire may be the more
advantageous of the two.
109
LITERATURE CITED
Balda, Russell P.
1975. Vegetation structure and breeding bird
diversity. In Proceedings of the symposium on
management of forest and range habitats for
nongame birds. p. 59-80. USDA For. Serv. Gen.
Tech. Rep. WO-1, Washington, D.C.
Canfield, R. H.
1941. Application of the line interception
method in sampling range vegetation. J. For.
39(4):388-394.
Daubenmire, Rexford.
1968. Ecology of fir in grasslands.
Ecol. Res. 5:209-266.
Advan.
Feist, F. G.
1968. Breeding bird populations on sagebrush-
grassland habitat in central Montana. Audubon
Field Notes 22:691-695.
Gashwiler, Jay S.
1977. Bird populations in four vegetational
types in central Oregon. USDA Fish. Wild. Serv.
Spec. Rep. , Wild. No. 205.
International Bird Census Committee.
1970. An international standard for a mapping
method in bird census work recommended by the
International Bird Census Committee. Audubon
Field Note 24(6) :722-725.
Johnson, Douglas H.
1976. Effects of grassland burning on breeding
birds — preliminary report. Proc. North Dakota
Acad. Sci. 30(1):24.
MacArthur, Robert H. , and John W. MacArthur.
1961. On bird species diversity. Ecology 42(3):
594-598.
Mcintosh, Robert P.
1967. An index of diversity and the relation of
certain concepts to diversity. Ecology 48(3):
392-404.
Shannon, C. E.
1948. A mathematical theory of communication.
Bell Syst. Tech. J. 27(7) :623-656.
110
The Effects of Prescribed Burning on Mule Deer in Lava Beds National Monument
Alice Purcell, Roger Schnoes, and Edward Starkey
ABSTRACT
In the winters of 1976 and 1977, research was conducted to determine the impacts of prescribed burning on
mule deer, Odocoileus hemionus hemionus, in Lava Beds National Monument. Visual observations, radio telem-
etry, and pellet-groups were used to examine deer distribution, seasonal movements, and winter food habits
of deer. Individual deer occupied site-specific home ranges in the Monument. Home ranges were not aban-
doned or extended as a result of burning. Bitterbrush, Purshia tridentata, was the most important browse
species to deer during the winter months. Utilization of green shoots increased, particularly in burned
areas, as the winter progressed. Deer increased use of burned portions of their home ranges in spring.
Due to warming trends in March, herbaceous vegetation will usually be available to does during the criti-
cal third trimester of gestation. Based on the size of home ranges, typical weather patterns, and the size
and mosaic nature of prescribed burns, deer probably will not be adversely affected by burns and may bene-
fit from them.
Alice Purcell, Roger Schnoes, and Edward Starkey,
Cooperative Park Studies Unit, School of Forestry,
Oregon State University, Corvallis, Oregon.
HI
Figure 1. — Location of Lava Beds National Monument,
California, and summer range of migratory deer.
INTRODUCTION
In the winters of 1976 and 1977, research was con-
ducted to determine the impact of prescribed burn-
ing on mule deer, Odocoileus hemionus hemionus , in
Lava Beds National Monument, California (Schnoes
1978). A second phase of the study was initiated
in the spring of 1977 and continued through the
spring of 1978. During the second phase, addi-
tional information was obtained on seasonal move-
ments and use of burns by deer that wintered in
the Monument. During the winters of 1976-78, data
were collected on distribution and home ranges of
deer in Lava Beds, movements and use of cover by
deer in response to weather, and food habits of
deer, especially as influenced by burns. Also, mi-
gration routes and summer range of migratory deer
were located during April and May, 1978.
STUDY AREA
Lava Beds National Monument is in northeastern
California, approximately 72 km south of Klamath
Falls, Oregon. It is part of the Basin and Range
Physiographic Province as described by Franklin
and Dyrness (1973). The climate is semi-arid with
a mean annual precipitation of 36.8 cm. Summers
are warm and dry with a mean maximum temperature
of 27 °C; winters are cool with an average minimum
temperature of -4.4°C. The Monument lies on the
northern flank of the Medicine Lake Highlands. The
land rises gradually from 1 250 m at the northeast-
ern boundary to 1 700 m in the southwestern corner
(fig. 1). Elevation continues to rise south of
the Monument to over 2 100 m in the Medicine Lake
Highlands.
The northern two-thirds of the monument is charac-
teristic of a shrub-steppe habitat (fig. 2). Domi-
nant shrubs are mountain big sagebrush, Art erne sia
tridentata ssp. vaseyana, rabbitbrush, Chrysotham-
nus nauseosus and C. viscidif lorus, and horsebrush,
Tetradymia canescens. Endemic bunchgrasses of
that area include blue bunch wheatgrass, Agropyron
spicatum, needlegrasses, Stipa thurberiana and S.
occidentalis, Sandberg's bluegrass, Poa sandbergii,
squirreltail, Sitanion hystrix, and Idaho fescue,
Festuca idahoensis. Prominent invader species are
cheatgrass, Bromus tectorum, and tumbling mustard,
Sisymbrium altissimum.
The southern one-third of the Monument is a dense
western juniper, Juniperus occidentalis, and curl-
leaf mountainmahogany , Cercocarpus ledifolius,
chaparral with an understory dominated by bitter-
brush, Purshia tridentata, and mountain big sage-
brush (fig. 2). The extreme southern portion
supports a ponderosa pine, Pinus ponderosa, forest
with white fir, Abies concolor , and incense cedar,
Libocedrus decurrens, and an understory of snow-
brush, Ceanothus velutinus , greenleaf manzanita,
Arctostaphylos patula, and bitterbrush.
The climate and topography of the summer range,
approximately 25 km south of the Monument, are
similar to that of Lava Beds. The mean maximum
temperature during the summer is 33.2°C; annual
precipitation averages 135.2 cm. Elevation rises
gradually into the Medicine Lake Highlands from
1 341 m on the plateau east of the Long Bell State
Game Refuge (fig. 1). Lower elevations support a
ponderosa pine forest similar to that in the south-
ern porton of Lava Beds. Higher elevations support
white fir and lodgepole pine, Pinus contorta.
METHODOLOGY
Pellet-group transects were established in 1962
and have been counted annually by Monument per-
sonnel. Each transect consisted of 10 circular
plots; each plot had a radius of 3.4 m and an area
of 0.004 ha. The transects were located through-
out the Monument (fig. 2). Additional transects
of the same design were established in 1976 in the
1973 Wildfire and the Fleener Chimneys and Hovey
Point prescribed burns (fig. 3). Those transects
were placed parallel to the edge of a burn and
spaced 150 m apart. Two transects were placed in
a burn, two were placed outside of a burn, and a
112
(•-'. •.-'.• I Sagebrush/
!■' ••■'••» cheatgrass
r5 f?l Shrub-steppe
Figure 2. — Vegetation zones of Lava Beds National
Monument, California. (Reprinted from Schnoes
1978).
middle transect was placed along the edge of a
burn. Counts from those transects established in
1976 provided data on the use of those areas by
deer before and after burning.
Food habits of deer during three winters were
obtained by a feeding frequency method. While ob-
serving deer, we also recorded plant species and
type of forage used by individual deer. For the
analyses of feeding data, the Monument was divided
into three zones (fig. 2); those zones reflected
the type of vegetation and forage available to
deer (Schnoes 1978). The southern zone roughly
coincided with the juniper-mahogany chaparral and
the coniferous forest. The 1973 Wildfire and the
1976 West Crescent and 1977 Caldwell prescribed
burns occurred in that zone (fig. 3). Juniper and
mountainmahogany occurred much less frequently in
the central zone, however bitterbrush and sagebrush
were common along with the bunchgrass (fig. 2).
The prescribed burn at Fleener Chimneys was con-
ducted in that zone (fig. 3). Sagebrush and rab-
bitbrush dominated the northern zone; cheatgrass
and tumbling mustard were also frequent (fig. 2).
The Hovey Point prescribed burn occurred in that
zone in 1976 (fig. 3).
Figure 3. — Location of burns in Lava Beds National
Monument, California.
During all three winters, deer were trapped in
Oregon single-gate traps or darted using Sernylan
(Phencyclidine hydrochloride), CI-744 (an experi-
mental drug), or a mixture of Rompun (Xylazine)
and M-99 (Etorphine). Deer were ear-tagged and 12
does were fitted with radio-collars in 1977 and
1978. Radio-collared does were monitored with a
truck-mounted null/peak antenna, or a four-element
hand-held yagi antenna when observation of an
animal was desired. The majority of data on home
ranges, seasonal movements and seasonal use of
burns by deer were obtained from eight radio-
collared does that were still being monitored in
the spring of 1978.
Home ranges were constructed according to the
minimum polygon method (Mohr 1947). A composite
home range was obtained for each radio-collared
deer from all of the animal's locations while on
the winter range. Seasonal home ranges were also
constructed. The seasons were delineated as:
June-August, summer; September-November, fall;
December-February, winter; and March-May, spring.
The chi-square statistic was used to test for sea-
sonal selectivity or avoidance of burns by deer,
in a fashion similar to that discussed by Neu et
al. (1974). Telemetry data for each animal were
analyzed by season.
113
Figure 4. — Locations of radio-collared deer on the
migration corridor south of Lava Beds National
Monument, California, during spring 1977 and 1978.
The chi-square statistic, two-way analysis of vari-
ance, and simple and step-wise multiple regression
were used in statistical analyses. The level of
statistical significance was set at p < 0.05 in all
analyses.
RESULTS AND DISCUSSION
Distribution and Home Range
Counts of pellet-group transects over an 18-year
period indicated that deer use during winter was
concentrated in the southern one-third of the Mon-
ument (Schnoes 1978). Based on those counts, and
an assumed defecation rate of 13 pellet-groups/day,
the number of deer in the Monument during winter
was estimated to vary between 1,500 and 2,500
Part of the deer population was non-migratory and
remained in the Monument during summer. Most of
those deer occupied the northern half where the
population level was low by comparison to that of
the migratory population.
Migratory deer arrived in the Monument in October-
December and migrated south from March-May. They
left in a south-southeasterly direction moving
along the lower hills of the Medicine Lake High-
lands (fig. 4). During the spring migrations in
1977 and 1978, most of the migratory deer delayed
briefly (usually less than 2 days) in "holding
areas", as defined by Bertram and Rempel (1977).
Distances travelled by deer to the summer range
varied from 1.6 to 65 km. It appeared that some
bucks moved to higher elevation in the Medicine
Lake Highlands and did not travel great distances.
Delays were also brief for most deer during the
fall migration. One doe, however, used two hold-
ing areas for 3 or 4 weeks each in the fall of
1977. She delayed in those same areas during her
1978 spring migration. Does that summered in or
near Long Bell State Game Refuge initiated their
fall migration early, probably due to a drought
that prevailed on the summer range for a 2d year
in 1977. Ashcraft (1961) documented an early fall
migration for deer that summered in the McCloud
Flats during the drought year of 1955.
Drought probably had the greatest influence on
movements of deer while on the summer range. Home
ranges of migratory deer were larger during summer
than at other seasons (does S-L, table 1) and aver-
aged 970 ha. In early August 1977, two does that
summered on the southern fringe of Long Bell State
Game Refuge moved to higher elevation in the High-
lands (does S and Y, fig. 5). They remained there
until September 25. They returned to the area
south of the State Game Refuge and were located in
the Monument on September 30. Since deer were mon-
itored for only one summer, it cannot be concluded
that such late summer shifts are not an annual oc-
currence. It is conceivable, however, that drought
induced early dry forage conditions at the lower
elevations, and that deer moved to higher eleva-
tions for forage in an earlier phenological stage.
In the fall of 1977, radio-collared migratory deer
returned to the same home ranges in the Monument
that they occupied in the winter and spring of
1977. Composite home ranges of radio-collared does
ranged from 287 to 5 022 ha (table 1); seasonal
home ranges of six does were largest in the fall
(averaging 1 424) and smallest in the winter
(averaging 671 ha).
Two-way analysis of variance showed no significant
difference in size of home ranges among seasons;
however, a significant difference occurred among
sizes of home range of individual deer. Deer in
the northern half of the Monument had significantly
larger (P < 0.05) home ranges than those of deer
in the southern half. That difference was attrib-
uted to disparity in the type of cover and forage
available to deer in the two areas. The juniper-
mahogany chaparral of the southern zone provided
dense cover for deer. In addition, bitterbrush
and mountainmahogany , two browse species selected
most by deer in all three winters, were abundant
in that zone. The shrub-steppe habitat of the
northern and central zones provided a more open
cover type for deer; bitterbrush and mountainma-
hogany were less abundant. Disparity in the deer
population levels between the two areas probably
was influenced by available cover and forage, and
in turn, may have had an interactive influence on
home range size.
All radio-collared deer used the same home ranges
while in the Monument in 1977 and 1978; however,
areas of concentrated use within individual home
ranges varied between years (fig. 6). Shifts in
114
Table 1 — Seasonal and composite home ranges in hectares of radio-collared does in Lava Beds National
Monument, Calif orniai/
Sampling
Composite
Deer
Summer
Fall
Winter
Spring
period
home range
DEER IN SOUTHERN
ZONE
A 1977
1542 (48)
—
605 (105)
369 (83)
1/77-4/77
659 (188)
1978
—
—
483 (41)
581 (88)
1/78-5/78
700 (129)
S 1977
584 (22)
491 (91)
—
152 (37)
3/77-5/78
726 (320)
1978
—
—
424 (72)
566 (109)
—
—
Y 1977
747 (27)
381 (96)
—
67 (27)
3/77-5/78
636 (301)
1978
—
—
346 (80)
581 (97)
—
—
L 1978
—
—
217 (82)
251 (120)
11/77-5/78
320 (235)
0 1977
289 (67)
186 (137)
—
354 (63)
4/77-5/78
463 (419)
C 1977
775 (34)
—
—
209 (23)
—
—
DEER IN CENTRAL AND NORTHERN ZONES
B 1977
406 (72)
1811 (121)
2227 (91)
2126 (132)
1/77-4/77
2873 (223)
1978
—
_ —
1026 (98)
1648 (72)
1/78-5/78
5/77-5/78
1917 (154)
2563 (402)
W 1977
747 (64)
4313 (121)
—
889 (83)
4/77-5/78
4022 (437)
1978
—
—
1 (95)
1287 (71)
—
—
J 1978
~ ~
__-
623 (71)
711 (81)
1/78-5/78
922 (152)
A'Figures in parentheses give the number of locations.
Figure 6. — Home ranges and centers of activity of
radio-collared does A and B during the winter-
spring period of 1977 and 1978.
Figure 5. — Movements and home ranges of radio-
collared does S, Y, and A on the summer range
south of Lava Beds National Monument, California,
summer 1977.
115
Table 2 — Pellet-group counts for transects estab-
lished by Schnoes (1978) in three major burns in
Lava Beds National Monument
Year
Burned
Edge
Unburned
Five-transect
total
1973 WILDFIRE
1976
36
35
22
39
31
163
1977
38
27
43
33
24
165
1978
32
26
31
19
36
144
FLEENER CHIMNEYS BURN
the burns. During high winds, deer were often ob-
served feeding or moving along the lee side of lava
ridges or in lava depressions that interrupted an
otherwise open terrain. Observations of deer in
1978 indicated that unless accompanied by high
winds, rain and snow did not cause deer to seek,
shelter when feeding in open areas.
Data obtained from radio-collared deer confirmed
such observations; no significant correlation
occurred between distances traveled by does and
temperature, wind, precipitation, or cloud cover
(R2 = 0.025, P > 0.05) in 1978. Results from
the analyses of deer movements in the winters of
1975 and 1977 were similar (Schnoes 1978). More
severe winter conditions might have revealed an
association between movement patterns and general
weather conditions.
19761/
0
2
0
1
1
4
1977
1
21
0
7
15
44
1978
7
10
1
2
0
20
HOVEY POINT BURN
19761/
0
2
3
2
2
9
1977
13
17
6
15
7
58
1978
1
0
3
0
0
4
1/ Counts made before burning.
centers of activity between years probably were a
result of the mild winter weather in 1978. Warm
temperatures and plentiful rain resulted in an
early greenup, particularly at the lower eleva-
tions in the north. Does in that area moved into
burns, cheatgrass meadows, and the fields of the
Tule Lake National Wildlife Refuge in mid-February.
Deer continued to use those areas heavily through
spring. The early and extended use of burns, mead-
ows, and agricultural fields in 1978 resulted in
the northward shift in centers of activity (doe B,
fig. 6).
Centers of activ
southern half of
in 1978 (doe A,
of southward tri
collared does of
year. Those tri
in centers of ac
ported that mule
Oregon, spent mo
during mild wint
ity of migratory does in the
the monument shifted to the south
fig. 6). Frequency and duration
ps to coniferous forest by radio-
the southern zone increased that
ps resulted in the southward shift
tivity. Leckenby (1968) also re-
deer on the Silver Lake Range,
re time in ponderosa pine habitat
ers.
Movements and Use of Cover in Response to Burns
The winters of 1976-78 were relatively mild. There
were few extended periods of below-freezing temper-
atures; snow accumulation remained minimal (below
18 cm). As a result, few opportunities arose to
monitor the response of deer to extreme weather.
During periods of inclement weather in all three
winters, deer sought shelter in winds greater than
16 km/hr. Deer used topographical irregularities
in addition to vegetation as wind breaks, particu-
larly in the more open shrub-steppe habitat and
Use of Burns 1976-78
There was a significant increase (P < 0.05) in the
number of pellet-groups in the Fleener Chimneys
and Hovey Point burns the 1st year after the fires
(table 2). In 1978, the number of pellet-groups
remained high in the Fleener Chimneys burn. A
significant decline occurred, however, in the num-
ber of pellet-groups immediately outside that burn
and in the Hovey Point area (table 2).
Such declines may have reflected a general disper-
sal of deer in both the Fleener Chimneys and Hovey
Point areas, resulting from the wet, mild winter
of 1978. Use within the Fleener Chimneys burn
possibly remained high as a result of the greater
availability of herbaceous vegetation present
within the burn. The Hovey Point area, however,
contained many open meadows with plant communities
similar to those in the Hovey Point burn during
the 2d year following the prescribed fire. In
addition, agricultural fields of the Tule Lake
National Wildlife Refuge lie immediately north.
Herbaceous vegetation was abundant throughout the
winter in all those areas. Groups of deer were
often observed feeding and moving slowly through
the Hovey Point burn towards the refuge in the
evenings.
Seasonal Use of Burns
All does, with the exception of doe L, selected
burns or used them in proportion to their avail-
ability (table 3). Deer increased use of burns in
the spring of 1978 from that of the winter. An in-
creased use of the 1978 Wildfire area by some does
in the southern zone was also apparent in the fall
(fig. 7). Increased use of burns in the fall and
spring resulted in a shift in seasonal centers of
activity toward burns (fig. 7).
The weather was exceptionally mild in 1978 while
deer were on the winter range. Herbaceous forage
was available throughout winter in meadows and
burned areas. In such areas, deer fed predomi-
nantly on herbaceous vegetation. Availability of
palatable herbaceous vegetation in burns appeared
to be the primary factor influencing concentrated
use of burns by deer, particularly in spring.
116
Table 3 — Chi-square analysesi' of seasonal use of burns, based on proportion of burn available in annual
home range of radio-collared does, Lava Beds National Monument
Season
Southern zone does
Central and northern zone does
Winter 1977
—
—
—
0.3 (0)
—
0.1 (0)
—
Spring 1977
35.3 (+)
76.1 (+)
6.2 (+)
0.1 (0)
—
0.9 (0)
78.6 (+)
Summer 1977
—
—
—
—
302.1 (+)
0.1 (0)
Fall 1977
21.6 (+)
121.8 (+)
157.5 (+)
1.1 (0)
22.1 (-)
65.4 (+)
93.7 (+)
Winter 1977-
78
1.9 (0)
4.8 (+)
107.8 (+)
0.5 (0)
7.9 (-)
53.5 (+)
23.1 (+)
Spring 1978
5.9 (+)
178.9 (+)
3.2 (0)
46.2 (+)
l/X^ > 3.84 are significant at P < 0.05; (+) indicates selection for burn, (-) indicates selection
against burn, and (0) indicates use of burn in proportion to its availability.
Feeding Habits
During all three winters, bitterbrush was the most
important browse species and represented 75 per-
cent of all browse feeding observations. Mountain-
mahogany was the second most important browse
species. In the northern zone, where bitterbrush
was sparse and grasses and forbs abundant, deer
diets consisted almost entirely of grass and forb
species. Deer in that zone often fed in the fields
of the Tule Lake National Wildlife Refuge and shrub
species did not appear to play an important role
in the diet.
In the southern and central zones, fire had the
greatest influence on plant composition and thus,
the availability of different forage classes. Ef-
fects of fire on the composition of deer diets were
most evident in those zones. In the northern zone
where the Hovey Point burn was conducted, cheat-
grass and tumbling mustard already were very abun-
dant; thus, rosettes and shoots of those species
were available without burning. All fires that oc-
curred in the southern and central zones resulted
in invasion of cheatgrass and tumbling mustard in
areas where those species were not very abundant
previous to burning.
Use of green forage throughout the Monument re-
mained high during all three winters (fig. 8).
The deer in the central zone gradually increased
use of green forage during winter. Low use of
green forage in February 1977, in contrast to the
other 2 years, was significant (P < 0.05) and prob-
ably was due to the very dry weather during that
month, which may have caused a delayed greenup of
grasses and forbs. That was particularly evident
in areas outside the burn (fig. 8).
In the 1973 Wildfire, use of green forage continued
to increase during February 1977 (fig. 8). Grass
and forb species in the unburned portions of the
central and southern zones may have had to compete
with the shrub species for the little moisture that
was available during that month. Inside the burn,
it was probable that low shrub densities reduced
that effect.
Figure 7. — Shifts in seasonal centers of activity,
summer 1977 to spring 1978.
117
SOUTHERN ZONE
^Ln„.^l
197} WILDE I RE
m
HEADQUARTERS BURN
CENTRAL ZONE
FLEENERS BURN
WEST CRESCENT BURN
jiMill
i
D<
D'
Figure 8. — Percent of relative use of green shoots
by deer in the winters of 1976-78, Lava Beds
National Monument.
In the southern zone, increased use of green forage
was less abrupt and reached a lower peak in March
than in the central zone. Again, the decline in
February may have reflected a lower production in
1977. In that zone, there were large areas of very
dense mountainmahogany and juniper stands, with
minimal grass and forb production in the understory.
The lack of availability probably was responsible
for continued high use of browse in late winter
(fig. 8).
In all the burns, relative use of green forage was
considerably higher than in surrounding areas.
Such use was extremely high in burns in the south-
ern and central zones during the winter that imme-
diately followed burning, despite the fact that
areas of unburned vegetation remained in those
burns. The second year following the Fleener Chim-
neys and West Crescent burns (1978), use of green
forage remained high. However, there was lower use
in January, as deer started to browse resproutlng
and unburned shrubs available in those burns (fig.
8).
Unlike other burns, the 1973 Wildfire was intense,
and relatively homogeneous. Increased use of green
forage was more gradual during 1976 and 1977
(fig. 8), probably due to a greater amount of snow
in early and mid-winter of those years. Snowfall
was high in February 1976 and January 1977. With
snow accumulation reducing availability of shoots,
deer in the 1973 Wildfire switched to the use of
dried tumbling mustard stems protruding above the
snow. Snowfall was low during winter 1978, but
precipitation was normal and came in the form of
rain. Green forage was available in that burn
throughout winter, and deer use remained high.
Effects of Burns on Home Ranges
Migratory and non-migratory does were not displaced
by recent burns within or adjacent to their home
ranges. Furthermore, they did not extend their
home ranges to include new burns. For example, the
prescribed headquarters burn, conducted the summer
of 1977, did not attract radio-collared does whose
winter home ranges were adjacent to it.
Doe 0, a non-migratory deer, had a small corner
of her home range burned in the 1977 headquarters
burn. Her use of the newly burned portion in-
creased significantly (P < 0.05) from her use of
that area during the previous year; however, she
did not extend her home range to include a greater
portion of the burn. Her spring home range of 1978
was much smaller than that in 1977. She restricted
her movements to the new burn and a small area sur-
rounding Monument headquarters that included part
of the 1973 Wildfire. Apparently the new burn
caused a reduction of her 1978 spring home range.
Radio-collared doe A exhibited a different response
to a similar situation. The 1977 Caldwell burn
affected a small portion (4 percent) of her 1978
composite home range. In contrast to doe 0, she
neither significantly increased or decreased her
use of the burned area from that of the previous
year. She used the burns within her 1978 compos-
ite home range in proportion to their availability
(table 3); her movements in 1978 seemed unaffected
by the presence of the new burn within her home
range.
Thus, deer showed a strong fidelity to their home
ranges in the Monument; the presence of burns did
not induce an extension or abandonment of those
home ranges. Migration data in the spring of 1978
suggested that deer also returned to the same hold-
ing area on transitional ranges and home ranges on
the summer range. Those data support Schnoes
(1978) hypothesis of site-specific home ranges for
individual deer.
Potential Impacts of Prescribed Burning on Mule
Deer in the Monument
Prescribed burns that have been conducted at Lava
Beds have not been detrimental to deer that winter
in the Monument. General observations and evidence
from four road-killed deer in the winter of 1978
indicated that animals were in excellent physical
condition during that year. In addition, two bucks,
which were harvested at 2.5 yrs of age during the
1977 hunting season, had four-point antlers. Such
antler growth implies that those deer were on a
high plane of nutrition. Mild weather prevailed
during the winter of 1977-78 and placed minimal
stress on deer during that period. Plentiful rain-
fall produced herbaceous growth in burns that re-
mained available through the fall and spring.
An examination of weather records for the Monument
over the past 18 years indicated that severe win-
ters were not frequent in Lava Beds. Winter
weather in 1976 and 1977 was milder than the 16-
year average, however it closely approached it.
During those winters, Schnoes (1978) reported that
deer "exhibited elasticity in response to the al-
tered environment in taking advantage of the bene-
fits of both burned and unburned areas." Even in
severe winters, herbaceous forage most likely would
become available in spring prior to migration, due
to warming trends in March.
118
The spring period, March-May, corresponds to the
final 3 months of fetal development in pregnant
does. Pregnancy demands additional energy and pro-
tein, particularly in the last trimester of gesta-
tion, because of rapid fetal growth rate (Moen
1973). Numerous studies have indicated that nutri-
tional deficiencies in does during the last tri-
mester of pregnancy result in increased neo-natal
mortality of fawns (Murphy and Coats 1966, Thompson
and Thompson 1953, Verme 1977).
Radio-collared migratory deer in Lava Beds re-
mained in the Monument through mid spring in 197 7
and through late spring in 1978. Thus, it appears
that migratory deer have access to herbaceous for-
age in burns during the initial half of the third
trimester of gestation. Availability of herbaceous
forage in spring may increase productivity and fawn
survival of deer in the Monument. Because deer re-
turn to site-specific home ranges in the Monument,
it is likely that prescribed burns will have the
greatest impact on those deer whose home ranges
overlap burned areas.
Moderately large burns, ranging from 400 to
34 800 ha, and numerous smaller fires have occurred
on the transitional and summer ranges of migratory
deer, concomitantly with prescribed burns in Lava
Beds. Migratory deer are on those ranges for the
last half of the final trimester of gestation and
during lactation. As succession proceeds in the
burns south of the Monument, a serai stage of opti-
mum cover and forage for deer will be reached. The
large area impacted by those burns probably will
affect a significant segment of the migratory deer
population at Lava Beds. Burns on the transitional
and summer ranges will have the greatest influence
on productivity and fawn survival 5 to 10 years
after they occur (Salwasser 1979).
An eventual increase in the deer population in the
Monument appears likely in the view of the pre-
scribed burns on the winter and early spring ranges
in the Monument, and wildfires on the transitional
summer ranges to the south. Therefore, periodic
counts from pellet-group transects already estab-
lished in the Monument should be continued. Pellet-
group counts will not provide an accurate estimate
of herd size, however information obtained from
them will reflect general population trends. Fur-
thermore, they will continue to provide information
on distribution of deer in the Monument.
Personnel of the California Department of Fish and
Game conduct annual herd composition counts in fall
(bucks: does: fawns) and spring (adults:f awns) . Such
counts provide an index to fawn survival and thus
herd recruitment and population trends. These will
provide valuable information on the response of the
migratory deer to changing conditions on the summer
range (fall counts) and on the winter range (spring
counts) and should be monitored closely.
rease in
opulation
n-migratory
1 when com-
An increase
, would
It appears
the northern
-group tran-
er in the
keep person-
of those
There is also the possibility of an inc
the non-migratory segment of the deer p
in Lava Beds. The present number of no
deer in the Monument is relatively smal
pared to the number of migratory deer.
in the non-migratory component, however
result in increased use of the range,
that most non-migratory deer remain in
half of the Monument. Counts of pellet
sects and herd composition counts of de
northern portion of the Monument would
nel informed as to the population trend
deer.
CONCLUSIONS
The prescribed burning program in Lava Beds is
still in an early stage. Relatively mild winter
weather occurred during the years of this study.
Severe winter weather may alter the behavioral and
movement patterns of Lava Beds deer. In a more
severe winter, herbaceous vegetation will not be
as available and the importance of thermal cover
will increase. The initial number of deer present
in a burn and the size, frequency, and nature of
the burns will bear directly on the response of
the population in the future. To date, prescribed
burns have left pockets or areas of unburned vege-
tation, and the size of burns has been smaller than
the composite home ranges of most deer monitored
in this study. Until an opportunity arises to ob-
serve response of deer to burns in a severe winter,
the present burning program can probably continue
without harm to the deer population. It may in
fact, be beneficial to the population given typi-
cal weather conditions and recent habitat altera-
tions that have occurred on seasonal ranges south
of Lava Beds.
LITERATURE CITED
Anderson, Charles A.
1941. Volcanoes of the Medicine Lake Highlands
California. Calif. Univ. Dep. Geol. Sci. Bull.
25:347-422.
Ashcraft, G. C.
1961. Deer movements of the McCloud Flats deer
herd. Calif. Fish and Game 47(3) :14 5-152.
Bertram, R. C. , and Rempel.
197 7. Migration of the North Kings deer herd.
Calif. Fish and Game 63(3) : 157-159.
Franklin, J. F. , and C. T. Dyrness.
1973. Natural vegetation of Oregon and Washing-
ton. USDA For. Serv. Gen. Tech. Rep. PNW-8.
Pac. Northwest For. and Range Exp. Stn. , Port-
land, Oreg. 417 p.
Leckenby, D. A.
1968. Influences of plant communities on win-
tering mule deer. W. Assoc. State Fish and Game
Comm. 48:201-208.
Moen, A. N.
1973. Wildlife ecology: An analytical approach.
W. H. Freeman and Co. 458 p.
119
Mohr, C. 0. Salwasser, H.
1947. Table of equivalent populations of North 1979. The ecology and management of the Devils
American small mammals. Am. Wild. Nat. Garden Interstate deer herd and its range. Ph.D.
37:223-249. diss., Univ. of Calif. 377 p.
Murphy, D. A., and J. A. Coates. Schnoes, R. S.
1966. Effects of dietary protein on deer. N. 1978. The effects of prescribed burning on mule
Am. Wildl. Nat. Res. Conf. Trans. 31:129-138. deer wintering at Lava Beds National Monument.
M. S. thesis, Oreg. State Univ. 71 p.
Neu, C. W., C. R. Byers, and J. M. Peek.
1974. A technique for analysis of utilization - Thompson, A. M. , and W. Thompson.
availability data. J. Wildl. Manage. 1953. Effect on mild yield of the ewe and growth
38:541-545. of her lamb. Br. J. Nutr. 7:263-274.
Verme, L. J.
1977. Assessment of natal mortality in upper
Michigan deer. J. Wildl. Manage. 41(4) :700-708.
120
Fire in the Forests of Mount Rainier National Park
Miles A. Hemstrom
ABSTRACT
Infrequent, catastrophic fires have been important forces in the forests of Mount Rainier National Park.
The effectiveness of topographic features as fire breaks, the relative fire resistance of forest habitat
types (Franklin et al. 1979), and the natural frequency of large fires in different habitat types are
examined. Ridges, valley bottoms, and lower slopes are effective fire breaks. High as well as cool, wet,
and low elevations and wet habitat types are relatively fire resistant. The same habitat types seem to
experience lower frequencies of large fires. This information should be useful in fire management planning
in a variety of ways.
Miles A. Hemstrom, Pacific Northwest Forest and
Range Experiment Station, Forestry Sciences
Laboratory, 3200 Jefferson Way, Corvallis, Oregon
97331.
121
Figure 1. — Outline map of Mount Rainier National
Park, Washington. Dashed line is approximate
upper tree line.
To a great extent, the rich mosaics of forests in
the Pacific Northwest reflect past fires. Fires
initiate succession and trigger redistribution of
plant and animal species. Fires respond to cli-
matic factors, local environment, and fuel loads.
Variable stages of forest recovery after fires
produce mosaics of stand ages, species diversity
and abundance, and forest structure. Patterns of
tree ages in Mount Rainier National Park's (MRNP)
relatively untouched forests reveal the presence
of past large fires. Understanding the way fires
have altered these forests is vital to understand-
ing their succession, architecture, and species
composition and distribution. In this paper, the
roles of topography and forest habitat types as
fire breaks and the effects of forest habitat type
on large fire frequency are examined.
The purpose of this study was to answer several
specific questions about the natural or pre-modern
man role of large fires in MRNP's forest ecosystems:
1. How effective are topographic features as fire
breaks?
2. How effective are different forest habitat
types as fire breaks?
3. Are there differences in the frequency of large
fires in different habitat types?
A. How can habitat types and topographic features
help fire management planning?
Though Mount Rainier imposes strong orographic ef-
fects on weather patterns in the Park, its climate
is typical of the western slope of the Cascades.
Summers are warm and dry. Winters are usually wet
and cool. July temperatures at Longmire and Para-
dise (837- and 1 682-m elevation) (fig. 1) average
16.2° and 12.1°C, respectively. January tempera-
tures average -1.1° and -2.9°C. Most of the aver-
age annual precipitation, 205 cm at Longmire and
269 cm at Paradise, falls in the winter, accumulat-
ing as deep snowpack at higher elevations. Winter
storms generally track from southwest to northeast;
and because Mount Rainier causes a lee-side rain-
shadow, the river drainages in the Park's southern
and western sectors receive more precipitation than
those in the north and east (fig. 1).
Franklin et al.JL' described the Park's forest
communities and habitats based on 497 sample plots
and extensive field reconnaissance. The final
classification defined 17 habitat and community
types ranging from low elevation, wet forests
through mesic to dry and high elevation forests.
Plant communities and habitat types were mapped on
a 1:50,000 scale whole-park USGS topographic map.
The classification and map were upgraded after
each of four field seasons.
The pattern of pre-modern-man fire in MRNP is
representative of much of the central and northern
west slope forests of the Washington and Oregon
Cascades. Though stand ages range from less than
50 years to over 1,000 years, the majority of for-
ests are over 350 years old (Hemstrom 1979). Large
fires are infrequent and holocaustic. This pattern
seems widespread as far south as the McKenzie River
in Oregon (Franklin et al. 1979). There is some
indication that large fires coincided with pro-
longed regional drought (Hemstrom 1979). Modern
man may have increased the frequency of large fires
during the late 1800 's and decreased it since 1900
(Hemstrom 1979).
TOPOGRAPHIC FEATURES AS FIRE BREAKS
This study relied on maps of fire boundaries, tree
ages from early serai trees, and forest habitat
and community types which were produced by earlier
studies in the Park (Hemstrom 1979) (see footnote
1). To examine the role of topography on fire
behavior, I measured and classified the lengths of
fire boundaries of recent, clearly defined burns
into seven topographic classes: ( 1) major and
secondary ridges, (2) upper slopes (within 120-m
elevation of a major ridge), (3) mid slopes, (4)
lower slopes (within 120-ra elevation of a major
valley bottom), (5) draws and valley bottoms, (6)
snow avalanche tracks, and (7) other.
i/Franklin, J. F. , W. H. Moir, M. A. Hemstrom,
and S. G. Lewis. Forest ecosystems of Mount Rain-
ier National Park. (Manuscript in preparation.)
122
Table 1 — Percent of total fire boundary in major
topographic units from six recent, large fires at
Mount Rainier National Park
Topographic
feature
Percent of total
fire boundary-'-
Major ridges unforested
Major ridges forested
Side ridges
Upper slopes
Mid slopes
Lower slopes
Valley slopes
Along fall line^
Other3
14
10
12
8
10
12
15
9
4
-*- Total fire boundary: 288 km.
^Fire boundaries running along the fall line of
an otherwise featureless slope.
^Includes benches at various slope positions and
snow avalanche tracks.
Since fire boundaries form when a fire dies, their
position on topographic surfaces may reflect the
effectiveness of topographic features as fire
breaks. Under ideal fire conditions, a large fire
burns over forested topography until it encounters
major fuel breaks. Under adverse weather condi-
tions, a large fire may stop without encountering
a break or upon encountering a minor break such as
a low, forested ridge. If fires spread and stop
without regard to topography as long as sufficient
fuel is present, fire boundaries should materialize
at random with respect to all but unforested topo-
graphy. The extent to which fire boundaries are
concentrated on certain topographic features indi-
cates departures of fires from random movement and,
therefore, the effectiveness of those features as
fire breaks.
Despite the fact that weather and chance play im-
portant parts in determining where fires stop, cer-
tain topographic features seem to act as effective
fire breaks (table 1). In fact, the two most im-
portant topographic fire breaks, ridges and valley
bottoms, together account for over half the total
length of fire boundary. Smooth slopes account for
about 30 percent of the total length. Proportion
of fire boundary increases downslope; a reasonable
result considering convective heat movement. Ridges
and valleys both are effective fire breaks because
they require vertical shifts of fire movement, of-
ten downhill or onto wetter or cooler sites. Only
a small portion of fire boundary was on open, fea-
tureless slopes, oriented along the fall line; a
condition indicating fires that stopped without
regard to topography.
Abrupt fuel changes over short distances would seem
to make avalanche tracks effective fire breaks, but
only a small portion of fire boundary was along
avalanche tracks. In many cases, fires seem to
have burned straight across large avalanche areas.
Avalanche tracks often increase in size or are re-
juvenated after a fire burns anchoring vegetation
(Winterbottom 1974, Hemstrom 1979). An avalanche
track which should have been an effective fire
break might not have been as large or even have
existed at the time a fire burned through the area.
HABITAT TYPES AS FIRE BREAKS
To study the role of forest habitat types as fire
breaks, I first superimposed a map of six clearly
defined, recent burns on a whole-park forest hab-
itat type map. Unfortunately, classifying fire
boundary by habitat type to indicate their fire
resistance presents several potential sources of
error. Some habitat types are usually restricted
to certain topographic features, compounding the
effects of habitat type and topography. Stand age
boundaries within a habitat type may also represent
changes in fuel loads. In addition, habitat types
represent vegetation potentials and not necessarily
the vegetation actually present. They are, there-
fore, more accurate predictors of environment than
existing fuel loads. Many of MRNP's forests, how-
ever, are over 350 years old and are, at least in
terms of fuels, similar to climax stands.
Another source of error is that the proportion of
the landscape occupied by particular habitat types
may influence the amount of fire boundary in each
habitat type, irrespective of fire resistance. To
correct for this bias to some extent, I divided
the percent of the total fire line in each habitat
type by the percent of the total area burned which
was the same habitat type. The habitat types were
ranked according to this calculation of relative
fire resistance. If a habitat ranked high, the
length of fire line relative to burned area of that
habitat was low. In other words, most of the fires
which burned into that habitat also burned through
it without stopping. The habitat type would there-
fore be relatively less fire resistant than others
which rank low. Some bias may be introduced by
using a ratio of a linear quantity to squared (area)
quantity which might change the ratio purely on the
basis of the study area size. The ranks of habitat
types according to relative fire resistance should
not be affected.
Another way to look at natural burning rates by
habitat types is to reconstruct episodes of fire
back in time and calculate burn rates by habitat;
hectares burned per year per hectare of habitat
type for fires larger than 100 ha. In previous
reconstrucion of fires at MRNP (Hemstrom 1979), I
used a somewhat arbitrary set of rules to define
old burn boundaries. For this study, I modified
the rules to better incorporate topographic fire
barriers. I then reconstructed fires to the first
significant topographic fire break in all direc-
tions. In general, the new reconstructions are
more conservative than the earlier ones.
123
Table 2 — Percent fire boundary divided by percent
burned area by habitat type for six recent, large
fires at Moun t Rainier Nati ona 1 Pa rk
Habitat type1
F/B^
Rank
Tsuga heterophylla/Achlys triphylla
Abies la s 1 oca rpa /Valeriana sltchensis
Abies a ma bills and Tsuga heterophy 11a/
Caultherla shallon
Abies amabilis/Xerophyllum tenax
Abies ama bills/ Berber! s nervosa
Abies amabills/Rubus laslococcus/
Rubus laslococcus phase
Abies amabl lis /Oplopanax horrldum
Abies amabllls/Tlarella unifoliata
Chamaecyparis nootkatensls/Vacclnrum
ovalif olium
Abies amabilis/Ery thronium montanum phase
Tsuga heterophy 11a /Polys tic hum muni turn
Abies amabilis/Menziesla f erruginea
Abies amabil is /Rhododendron albif lorum
0.45
1
0.49
2
0.68
3
0.70
4
0.84
5
1.00
6
1.04
7
1.32
8
1.81
9
2.00
10
2.04
11
2.35
12
3.17 '
13
^Some habitat types are minor and not Included. (See footnote 1
In text. )
^F/B Is the percent of measured fire boundary in a habitat type
divided by the percent of the total burned area in the same habitat
type. Total fire boundary measured: 237 km. Total burned area:
14 368 ha.
Comparing the ratio of percent burn line to percent
burned area and reconstructed large fire frequency
allows a relatively independent evaluation of the
role of fire resistance and frequency by habitat
type. The first method does not rely on reconstruc-
tions but is limited to recent, well defined fires.
The second method reaches much farther back in
time but requires accurate reconstructions. Both
approaches depend on accurate habitat type maps and
tree age data.
Some habitat types appear to be significant fire
breaks (table 2). The Abies amabilis/Rhododendron
albif lorum (Abam/Rhal) and Abies amabilis/Menziesla
f erruginea (Abam/Mefe) habitat types appear to be
very fire resistant. Since both of these types
tend to occur on north facing, high elevation, wet
slopes, much of their fire resistance may be due
to topographic position. Sites where these types
occur also experience heavy snowpacks, and fuels
would be wet much of the time.
At the other extreme, the Tsuga heterophylla/Achlys
triphylla (Tshe/Actr), Abies lasiocarpa/Valeriana
sitchensis (Abla/Vasi), Tsuga heterophylla and
Abies amabilis/Gaultheria shallon (Tshe/Gash and
Abam/Gash), Abies amabilis/Xerophyllum tenax (Abam/
Xete) and Abies amabl lis/Berberis nervosa (Abam/
Bene) habitat types all seem to burn readily in
large fires. The Abam/Xete and Abla/Vasi habitat
types tend to occur on high elevation, exposed
sites which are subject to summer lightning. They
are also dominant habitat types in the White River
drainage, the driest in the Park. The Tshe/Gash,
Abam/Gash and Abam/Bene habitat types are typical
of dry sites where fuels might be flammable early
in the season. The low fire resistance of the
Tshe/Actr habitat type may be an artifact of its
small extent relative to the other habitat types.
FIRE FREQUENCY FROM RECONSTRUCTED FIRE HISTORY
Analysis of topographic features as fire breaks
provide a basis for defining rules for reconstruct-
ing fire patterns in MRNP to 750 years ago. Calcu-
lations of burn rate for large fires and natural
fire rotation for the Nisqually, Ohanapecosh, and
White River drainages reveal a pattern of large
fire frequencies which change between habitat types
within a drainage and within habitat types between
drainages (table 3). The overall pattern of large
fire frequency in different habitat types strongly
resembles the pattern of fire resistance by habitat
types.
Over the three drainages, the Abla/Vasi, Abam/Bene,
Tshe/Actr, Abam/Gash, Tshe/Gash, and Abam/Xete
types experience large fires most often. The
Abies amabilis/Ery thronium montanum (Abam/Ermo) ,
Abies amabilis and Tsuga heterophylla/Oplopanax
horridum (Abam/Opho and Tshe/Opho), and Abam/Rhal
habitat types burn least often. Except for Abam/
Rhal, these were not the most fire resistant
(table 2). This difference may stem from my use
of the whole Park in fire resistance analysis but
only three drainages in fire frequency analysis.
The overall order of habitat types by fire resist-
ance and decreasing fire frequency is remarkably
similar, however.
There are some interesting, and in some cases unex-
plained, differences in fire frequency by habitat
type between drainages. The Abam/Mefe habitat type
ranks first in fire frequency in the Nisqually
drainage and sixth in both the Ohanapecosh and
White River drainages. The Abam/Xete habitat type
ranks second in the Nisqually, third in the Ohan-
apecosh and tenth in the White River drainage.
These examples may indicate gaps in our understand-
ing of successional status of some of the Park's
habitat types or may reflect bias from the rela-
tively few fires sampled.
In general, the fire frequency of the same habitat
type increases from the Nisqually to the Ohanape-
cosh to the White River drainage. The average
fire frequency increases and natural fire rotation
decreases from 0.0023 ha ha~l year--'- (438
years) to 0.0031 ha ha~l year--*- (324 years)
from the Nisqually to the White River drainage.
The average fire frequency and natural fire rota-
tion for the three drainages is 377 years, lower
than the whole Park natural fire rotation of 465
years (Hemstrom 1979).
APPLICATIONS TO FIRE MANAGEMENT PROBLEMS
This analysis provide
tion useful in: (1)
tive places to put fi
operations, (2) locat
frequency and natural
let-burn management f
areas where fires, ei
are frequent and nece
s fire managers with informa-
pointing out the most effec-
re lines during fire fighting
ing areas where natural fire
fire breaks might make
easible, and (3) indicating
ther prescribed or natural,
ssary parts of the landscape.
124
1 2
Table 3 — Fire frequency (FF) and natural fire rotation (NFR) by habitat type for the Nisqually.
Ohanapecosh, and White River drainages, Mount Rainier National Park
Habitat type-^
N
isqually
Oha
napecosh
White
d
rainage
drainage
drainage
Average
NFR
FF
Rank
NFR
FF
Rank
NFR
FF
Rank
NFR
FF
Rank
NA
NA
NA
191
.0052
1
282
.0035
3
275
.0035
1
400
.0025
4
303
.0033
3
258
.0039
2
295
.0034
2
NA
NA
NA
360
.0028
5
208
.0048
1
308
.0033
3
325
.0031
2
305
.0032
4
305
.0033
3
313
.0032
4
348
.0029
3
273
.0036
2
490
.0020
9
323
.0031
5
269
.0037
1
396
.0025
6
335
.0030
5
343
.0029
6
667
.0015
10
400
.0025
7
318
.0031
4
367
.0027
7
466
.0021
6
405
.0025
8
421
.0024
8
426
.0023
8
435
.0023
5
NA
NA
NA
NA
NA
NA
435
.0023
9
523
.0019
8
494
.0020
9
365
.0027
6
474
.0021
10
663
.0015
9
551
.0018
11
395
.0025
7
478
.0021
11
503
.0020
7
700
.0014
12
NA
NA
NA
535
.0019
12
729
.0014
11
510
.0020
10
592
.0017
10
616
.0016
13
Abies lasiocarpa /Valeriana sitchensis
Abies amabilis/Berberis nervosa
Tsuga heterophylia/Achlys triphylla
Abies amabilis and Tsuga heterophylla/
Gaultheria shallon
Abies amabilis/Xerophyllum tenax
Abies amabilis/Menziesia f erruginea
Abies amabilis/Rubus lasiococcus/
Rubus lasiococcus phase
Abies amabilis/Tiarella unif oliata
Tsuga heterophylla/Polystichum muni turn
Abies amabilis/Vaccinium alaskaense and
Chamaecyparis nootkatensis/Vaccinium
ovalif olium
Abies amabilis/Rhododendron albif lorum
Abies amabilis /Oplopanax horridum
Abies amabilis/Erythronium montanum
■'•Burned hectares per hectare of habitat type per year for fires over 100 ha.
^The time required to burn an area equal to the total area of each habitat type given its burn rate (Heinselman
1973).
•'The areal extent of some habitat types was insignificant. These are omitted. (See footnote 1 of text.)
The most effective places to construct fire lines
are located where natural topographic or vegetative
fire breaks would complement artificially decreased
fuel loads. It might be important not only to con-
struct a fire line on a ridge top, as is common
practice, but to place the line next to a naturally
fire resistant Abam/Rhal stand. While it might be
essentially useless to put a fire line through a
dense Abla/Vase, Abam/Xete, Abam or Tshe/Gash or
Abam/Bene stand, a similar fire line next to Abam/
Rhal, Abam/Ermo, Abam/Mefe, Abam/Vaal or Abam/Opho
habitat types could prove effective. There is no
assurance, however, that fires will stop on ridges
or in an Abam/Rhal stand if conditions for fires
are favorable.
Another application is in planning fire management.
In some areas, natural flammability and vegetative
and topographic fire breaks provide conditions
which might make possible let-burn fire manage-
ment. A slope covered with Abam/Bene habitat type
and bounded on three sides by ridges and fire re-
sistant habitat types might be left to burn if
weather conditions were suitable. But fires in
dense Abla/Vasi stands on gentle slopes abutting
valuable timber land outside the Park could easily
escape. A whole Park map could be divided into
units for let-burn management based on habitat
type, vegetative and topographic fire barriers,
and consideration of adjacent ownership.
Finally, fire frequency information by habitat
type indicates where fires are most important as
natural processes and where they are most likely
to occur in the future. The former could provide
a scale for evaluating the urgency of applying
prescribed burns or let-burn management. The
latter could be important to fire lookouts, to
fire bosses in action, and for planning future
developments such as trails or facilities.
CONCLUSIONS
1. Certain topographic features serve as effective
fire breaks, especially ridges and valley bottoms.
2. Certain habitat types are more fire resistant
than others. This may reflect their characteristic
topographic location and environment.
3. Certain habitat types experience natural, large
fires more frequently than others.
4. Both fire resistance and fire frequency vary
between habitat types within river drainages and
within habitat types between drainages. The driest
river drainages have the highest natural frequency
of large fires.
5. Natural vegetative and topographic fire breaks
and differential fire frequencies between habitat
types have important applications in fire manage-
ment.
125
LITERATURE CITED Hemstrom, M. A.
1979. A recent disturbance of forest ecosystems
Franklin, J. F. , A. McKee, F. J. Swanson, J. Means, at Mount Rainier National Park. Ph. D. thesis,
and L. Brown. Oreg. State Univ. , Corvallis. 67 p.
1979. Age structure analysis of old-growth
Douglas-fir stands: data versus conventional Winterbottom, K.
wisdom. Bull. Ecol. Soc. Am. 60:102. (Abstr.). 1974. The effects of slope angle, aspect and
fire on snow avalanching in the Field, Lake
Heinselman, N. L. Louise, and Marble Canyon region of the Canadian
1973. Fire in the virgin forests of the Bound- Rocky Mountains. M.S. thesis, Univ. of Calgary,
ary Waters Canoe Area, Minnesota. Quat. Res. Alberta.
3:329-382.
126
Forest Dynamics and Fuelwood Supply of the Stehekin Valley, Washington1
Bruce C. Larson and Chadwick Dearing Oliver
ABSTRACT
Stehekin Valley National Recreation Area in the North Cascades National Park complex of Washington is an
isolated valley used by permanent residents for fuelwood (and other uses) and by seasonal recreationists.
Upland forests of the Recreation Area consist of Douglas-fir, ponderosa pine, lodgepole pine, bigleaf
maple, and other species. These forests at first appeared to be perpetuated by all-age succession.
Examination, however, showed the forests exist in age classes which begin following large disturbances
such as fire (in 1889). Small disturbances such as selection cuttings do not allow new stems to be
recruited. The valley appears to go through large disturbances on 90- to 100-year cycles. Cuttings for
firewood should mimic these disturbances if the natural forest patterns are to be maintained.
-'•This study was performed through the University
of Washington, College of Forest Resources/U. S.
National Park Service Cooperative Park Studies
Unit (Contract No. CX 9000-9-0088).
Bruce C. Larson, research assistant, and Chadwick
Dearing Oliver, assistant professor, College of
Forest Resources, AR-10, University of Washington,
Seattle, Washington 98195.
127
X
A
Cascade
Ridge
V
\ Ferry
J Landing
Area
State of
Washington
Figure 1. — Location of study area in Stehekin,
Washington.
INTRODUCTION
The Stehekin Valley in the North Cascades National
Park complex of Washington State is unique for a
combination of geographic, social, and ecologic
factors. It is a National Recreation Area (NRA)
administered by the National Park Service, and
management other than preservation is mandated.
The Cascade Range and 55-mile (89-km) Lake Chelan
provide isolation and difficult access (fig. 1).
Outstanding beauty and vegetation diversity bring
more than 15,000 visitors each year, although
there are fewer than 130 winter residents.
The overwintering population has been expanding,
and the winter residents heat with firewood cut
from the NRA (to date, primarily dead wood has
been taken) and private lands within the valley.
The National Park Service managers needed
information about the long range firewood
availability and how they could properly balance
the valley's natural beauty, firewood cutting, and
winter population level.
Tne purposes of the study were to determine the
potential firewood supply and how and where it
could be cut. The objectives were accomplished by
examining the forested areas to determine the
standing volume of wood and to determine the stand
dynamics--the stand development pattern — of each
area.
This paper reports preliminary findings of forest
stand dynamics and their importance for management
considerations in one of four subareas of the
Stehekin Valley. The paper describes the natural
stand dynamics of the existing forest and the
effects and potential effects of various cutting
methods on the forest's development. The area is
divided into forest types based on uniform soil
substrate and the diameter, age, and species
distribution. The disturbance history of each
area is described.
LITERATURE REVIEW
Conventional forest inventories (Dilworth 1977,
Husch et al. 1972) have often been used to
determine stand volumes and growth rates. When
such inventories have been used for growth rate
predictions, a pattern of natural growth — either
all-age or age classes following large
disturbances — has been assumed.
Certain characteristics have been assumed to be
indicative of an all-aged forest — the "reverse
J-shape" frequency distribution of diameters, and
vertical crown stratification by species (Rough
1932, Meyer and Stevenson 1943, Braun 1950,
Philips 1959, Daubenmire 1968, Minckler 1974).
All-aged forests have been assumed to develop
naturally by the smaller (and presumably younger)
stems gradually replacing the larger (and
presumably older) stems. Selective cutting
practices in such stands where few larger trees
are taken at a time have been assumed to mimic the
natural growth pattern.
Studies in other forest types in North America
(Hough and Forbes 1943, Henry and Swan 1974,
Oliver 1978, Stubblefield and Oliver 1978, Wierman
and Oliver 1979) have examined the age structures
and growth patterns in forests with diameter and
stratification characteristics previously attri-
buted to all-age stands. These characteristics
were found to exist in stands in which all stems
originated in a distinct time interval (age class)
following a large disturbance. A stand could have
more than one age class where more than one large
disturbance occurred but did not eliminate all
previously standing trees. Small disturbances
(such as the removal of a few trees) did not
create a new age class, but allowed the existing
stems to increase in size (Oliver and Stephens
1977).
Forests in more mesic regions than the Stehekin
Valley have been described as having periodic
disturbance cycles (Loucks 1970, Heinselman 1973,
Wright 1974) , in which the forests became more
susceptible to large disturbances (such as fire
following litter buildup) as they grow old. It
was not known if a similar disturbance pattern
occurred in forests of the types found in the
Stehekin Valley.
The patterns of development and management of
stands similar to those in the Stehekin Valley
have been of concern (USDA Forest Service 1978).
It was unknown if the forests in the Stehekin
Valley developed in an all-aged pattern or in an
age-class pattern following large disturbances.
128
Management by selective cutting assumes the
all-aged pattern of stand development can occur.
Management by more intensive cutting assumes new
stems begin after large disturbances mimicked by
the cutting. Conditions following each type of
cutting may be different in several ways:
(1) Successional pattern: One cutting pattern
would follow the natural stand dynamics and
produce stands similar to those naturally there;
the other cutting pattern could produce forests of
quite different species and appearances.
(2) Volume growth: The tree growth for firewood
may prove quite different if the stands were
managed on an age-class basis or on an all-age
basis.
(3) Regeneration: For forests of the same type
to be perpetuated in the valley, any cutting must
mimic the natural stand development pattern enough
to allow trees to regenerate.
It is important, therefore, to understand how each
forested type develops — either in age classes
following large disturbances or by all-aged
gradual stem recruitment — within the valley if the
forests are to be manipulated and perpetuated for
natural beauty, recreation, and firewood uses.
PROCEDURES
Study Area
The study area was approximately 4,000 acres,
(1 600 ha) about 15 percent of which is private
ownership. It could be divided roughly into four
forested subgroups: river floor plain, moist
alluvial, upland, and steep sidewall forests.
This paper describes the results from the upland
(nonf looding) subgroup — a highly accessible area
and therefore desirable for firewood cutting.
The valley is typed as Abies grand! s and
Pseudotsuga menziesii zones by Franklin and
Dyrness (1973). The dominant tree species are
Douglas-fir, Pseudotsuga menziesii [Mirb. ] Franco,
and ponderosa pine, Pinus ponderosa Dougl. ex Laws.
The flood plain areas also have cottonwood,
Populus trichocarpa Torr. and Gray, bigleaf maple,
Acer macrophyllum Pursh, red alder, Alnus rubra
Bong. , grand fir, Abies grandis [Dougl. ex D. Don]
Lindl. , and western redcedar, Thuja plicata Donn.
ex D. Don.
Miners and trappers frequented the valley in the
19th century using some wood for houses and fuel.
Regular steamboat service and tourist trade from
Chelan began in 1888. Fuelwood for the boat was
cut in the lower portion of the study area in
contiguous patches that eventually resembled
clearcuts. A large, hot fire swept the valley in
1889. After 1900, selective cuttings were done
continuously in many areas. Larger fire remnant
trees were removed for housebuilding and small
trees were thinned for firewood.
The U.S.D.A. Forest Service managed much of
Stehekin Valley until the late 1960's, at which
time management was assumed by the National Park
Service.
This glacial valley is typically U-shaped, 1 to 2
miles (1.5 to 3 km ) wide, with steep side walls.
The glacial drift has been resorted by the Stehekin
River. Upland stands lie on river terraces indi-
cating a once large and shifting glacial river.
Several creeks originating from large empty cirques
enter the valley from the side walls. The creeks
moved much glacial material into raised alluvial
fans of loose rubble on the valley floor when the
mountain glaciers were retreating. The streams
now are fed by snowmelt and no longer are powerful
enough to flow over the large fans. They flow to
one side before joining the Stehekin River.
Elevation of the study area ranged from 1,140 to
2,463 ft (348 to 751 m) . Upland areas studied in
this paper were located between 1,300 and 1,600 ft
(396 and 488 m) . Rainfall ranges from 34 to 40 in
(86 to 102 cm) (lowest in southeast) — most falls
as snow in winter; summers are very dry. Average
temperature is 49°F (9°C). On the average, the
maximum temperature is below 32 °F (0°C) during 49
days each year and the maximum exceeds 95°F (35°C)
during 9 days (Donaldson and Ruscha 1975).
Field Procedures
The valley was studied during the summer of 1979
by stratified sampling techniques. The area was
mapped into stand types (Dilworth 1977) using:
1978 true color aerial photographs; interviews
with residents and National Park Service personnel
for disturbance information; and visual inspection
noting tree species, lesser vegetation, evidence
of disturbance, and information from soil pits.
Variable size (prism) plots were established in
each type using a 10-f actor prism. Trees 7.0 in
(17.8 cm) or less in diameter at 4.5-ft (1.4-m)
d.b.h. were sampled using a fixed plot 30 ft
(7.2 m) in radius. Data were collected in one of
two ways. "Simple" plot data included: diameters
and subjective crown classes (Smith 1962) of
trees, representative codominant heights, notes of
lesser vegetation; and evidence of past
disturbance including charcoal, fire scars, wind
breakage or blowdown, and stumps, and notes of
evidence of past forest stand structure including
down wood and litter searches. "Complete" plots
included all above information plus the heights of
all trees and increment cores of two trees from
each 2 in diameter class, if possible. Increment
cores were brought back to the laboratory and their
ages determined using a dissecting microscope.
Data Analysis
Stems per acre were calculated by species and
diameter class. Averages of four or more plots
were used in each type. Age structure of stands
was determined. Proportion of trees sampled in
each 10-year age class was calculated by species.
RESULTS AND DISCUSSION
The upland, accessible forest subgroup described
in this study was subdivided logically Into four
types based on uniformity of soils. Character-
istics of each area are shown in table 1 and are
discussed below.
129
Table 1 — Characteristics of upland, accessible forest types in Stehekin Valley
Type
Tree species
Area
(approx. )
Soil
Disturbances
Diameter
range
Age range
(years)
Peak age
classes
(years)
Moist
conifer
stand on
lower
terraces
Transition
zone between
fans and up-
land terraces
Glacial
drainage
fans
Douglas-fir
bigleaf maple
ponderosa pine
Douglas-fir
bigleaf maple
Douglas-fir
ponderosa pine
bigleaf maple
5 acres excessively well- extensive fire
(2.0 ha) drained glacial 1889
rubble, much
sorting much cutting
prior to 1900
30 acres excessively well- extensive fire
(12.1 ha) drained glacial 1889
rubble
250 acres excessively well- extensive fire
(101.2 ha) drained glacial 1889
rubble, many large
boulders numerous small
fires
1 - 44 in
(2.5 - 1118 cm)
1 - 44 in
(25 - 111.8 cm)
1 - 40 in
(2.5 - 101.6 cm)
40-170
50-180
30-375
65
75, 155
65, 155, 255
4. Upland
terraces
5. Upland
terraces
a. uncut
Lodgepole pine
Douglas-fir
bigleaf maple
willow
Douglas-fir
ponderosa pine
bigleaf maple
b. partially
cut
dogwood
75 acres
(30.4 ha)
175 acres
(70.8 ha)
75 acres
(30.4 ha)
excessively well-
drained glacial
rubble, some
sorting
excessively well-
drained glacial
rubble, some
sorting
extensive fire
1889
extensive fire
1889
some cutting in
4b, occasional
snowslides
1 - 40 in
(2.5 - 101.6 cm)
1 - 40 in
(2.5 - 101.6 cm)
1 - 42 in
(2.5 - 106.7 cm)
50-130
85
30-150
30-80
65, 125
65
Moist Stand Type
The type consisted of excessively well-drained
material; moistness resulted from close proximity
of water table. This area was cut over prior to
1900 and burned in 1889. Some burned wood was
collected for steamboat fuel after the fire.
Figure 2 shows the species distribution by dia-
meters and ages. The large range of diameters and
" re verse-J- shape" diameter distribution showing
many small stems and few large ones can give the
appearance of an all-aged stand, with the smaller
trees gradually replacing the larger (and presum-
ably older) ones. The age distribution, however,
shows that the stems invaded during an interval
after 1889, the time of the large fire. As has
been found in more mesic forests in other parts of
North America, stems have not been continuously
invading the area, but invaded in a distinct wave
following the disturbance.
Transition Zone Type
The type represented a transition between moist
stands and extremely dry glacial alluvial fans.
This area was not cut over, but selective cutting
was a continual influence.
Figure 3 shows a similar diameter distribution to
the moist type. The age structure shows two age
groups, or classes: one after the fire of 90 years
ago and one approximately 150 years old.
130
□ DOUGLAS -FIR
IHD BIGLEAF MAPLE
Q PONDEROSA
60r
(/)
UJ
UI
or
i-
u.
O
or
LJ
CD
ID
5 15 25 35
DIAMETER (AT 4.5 FEET), INCHES
15 r-
10 -
5 -
I I I ffTil
I I I I I I I I iFil I I
25 55 85 115 145 175
AGE BY 10- YEAR PERIODS
Figure 2. — Diameter and age distributions by
species for "Moist conifer stand on lower
terraces" type.
UJ
or
o
<
or
UJ
0-
cn
UJ
UJ
or
40
20
0
EB BIGLEAF MAPLE
□ DOUGLAS -FIR
lllt-m+m
i rrli h li rru
5 15 25 35
DIAMETER (AT 4.5 FEET), INCHES
15 r
ui
ui
P 10
or
ui
CO
25 55
AGE BY 10
85 115 145 175
YEAR PERIODS
Figure 3. — Diameter and age distributions by
species for "Transition zone between fans and
upland terraces" type.
Glacial Fan Type
This type was also low elevation, but driest
because the glacial drift consisted of large rocks
and was excessively well drained. Dryness resulted
in aggregated and somewhat open stands. Lesser
vegetation was primarily grasses, Arctostaphylos,
and Ceanothus.
Figure 4 shows a strong "reverse J-shape" diameter
distribution for each species as well as the com-
bination of species. Age structure shows three
age groups, or classes. The age distribution
within each "group" is quite broad, probably
because it took several decades for trees to grow
large enough to exclude later arriving ones.
Examination of a small recent burn (1967) on a
similar soil showed new tree stems were still
invading the area in 1979.
Many trees remain in the pre-1889 age groups,
probably beause the lack of fuels on this dry site
kept the fire incomplete.
Upland Terraces with Lodgepole Pine Type
The type was located on higher, sandy terraces
with north aspects. Dense, small diameter stands
consisted primarily of lodgepole pine with a large
component of Douglas-fir. Stands average over 195
trees per acre of 6- to 8-in (15.2- to 20.3-cm)
d.b.h. and 130 trees of 4- to 8-in (10.2- to 20.3
cm) d.b.h. All trees over 12-in (30.5 cm) d.h.b.
were Douglas-fir.
The age distribution was very narrow. Sixty
percent of trees whose ages were determined (all
species) were aged 70 to 90 years. Over 90
percent of lodgepole pines were 70 to 90 years.
The pine component was post-1889 (fire) origin.
The stands were located in a cold air drainage,
which may partally explain the lodgepole pine.
The terraces showed similar age and diameter
distributions to other types but with different
species.
131
60
540
<
or
Ul
Q_
[520
UJ
or
rl
:|i
0
i It, li li
D DOUGLAS-FIR
E3 FONDEROSA PINE
d BIGLEAF MAPLE
J_L
'''''
5 15 25 35
DIAMETER (AT 4.5 FEET), INCHES
u
or
<
or
UJ
o.
en
ui
UJ
or
t-
NO CUTTING
PARTIAL CUTTING
8 16 24 32 40
DIAMETER (AT 4.5 FEET), INCHES
I5r
en
ui
&I0
i-
u.
o
oS 5
3
0
I I n ll l! h hi ll I
m±m
LiLiLli
* nn i ■ rrii fTii
Figure 4. — Diameter and age distributions by
species for "Glacial drainage fans" type.
25 55 85 115 145 75 205 235
AGE BY 10-YEAR PERIODS
u_ <=> 100%
50% -
8«
h en
or
o en
Q. ui
O Ld
or or
o. i-
40 80 120 160 200
AGE BY 20 -YEAR PERIODS
Figure 5. — Diameter and age distribution of all
species for "Upland terraces" type, comparing
cut and uncut stands.
Upland Terraces of Douglas-fir
The type included most upland terraces in the
valley and consisted of two age groups. Some
stands had never been cut, others had been
partially cut (most about 20-30 years ago). The
cutting was selective removal of some large fire
remnant trees and some thinning of other, smaller
diameter stems, primarily for firewood. Many
stumps were 10- to 14-in (25- to 35-cra) diameter
and some exceeded 20 in (50.8 cm). The cut stands
had essentially returned to crown closure.
Figure 5 shows the diameter and age distribution
of both cut and uncut stands of this type. Again,
a distinct peak of ages is evident indicating a
wave of regeneration after the 1889 fire and an
exclusion of later stems. No fires had occurred
in this area to eliminate possibly younger trees,
apparently the selective cutting done in these
stands was not a suitable disturbance to allow new
stems to become established. This indicates that
it may be unfeasible (as well as unnatural) to
perpetuate the stand in an all-aged condition.
GENERAL DISCUSSION AND CONCLUSIONS
Stand Dynamics
The age distribution for all stands is shown in
figure 6. Distinct age classes, indicating waves
of regeneration, are apparent. As a whole, the
area was multi-aged but not all-aged.
The greatest proportion of trees originated in a
wave beginning after the last documented major
disturbance (fire of 1889). This wave is also the
only one still to contain nonf ire-resistant species
(such as hardwoods and lodgepole pine).
Although the stands originat
classes, they had an all-age
the diameter distribution,
also the result of vertical
(Smith 1962). This age clas
found elsewhere (Hough and F
Stephens 1977, Oliver 1978,
Oliver 1978) but has been ha
east Cascade slope stands,
ment time following the dist
data have misled observers t
stands are all-aged.
ed in distinct age
appearance because of
This appearance is
crown stratification
s structure has been
orbes 1943, Oliver and
Stubblefield and
rd to identify in dry,
A long stem recruit-
urbance and a lack of
o conclude that the
132
Previous waves of regeneration may have been
coincidental with either major disturbances (such
as fire or perhaps catastrophic insect defoli-
ations) or favorable climatic conditions during
good seed years.- Several factors suggest fire was
the cause: (1) favorable climate and seed crop
conditions probably would have lasted for a
shorter time; therefore the age range within each
"class" would not have been as broad; (2) fires
have been a factor in the valley for a long time;
buried soil horizons were found to contain
charcoal.
A fire cycle dependent on the buildup of fuel such
as has been found in less dry forests of the Mid-
west (Heinselman 1973) probably exists. Figure 6
indicates a possible 90- to 100-year cycle of fire
in the valley. At present, 90 years after the
fire of 1889, a large fuel buildup again presents
a fire danger.
The partially cut stands shown in figure 5
indicate that regeneration does not follow. minor
disturbances (no trees less than 25 years old were
found here). Peaks of regeneration following
major disturbances and no regeneration following
minor ones also have been documented in the New
England forests (Oliver and Stephens 1977).
MANAGEMENT IMPLICATIONS
The findings to date are preliminary. More
analyses will be made of the data collected;
however, the study appears to contain several
important management implications:
1. Forest fires may be a real hazard in the
valley at present because fuel has accumulated
since the last cyclic fire. Past natural disturb-
ances were widespread and catastrophic. It is
possible that the fuel buildup rate exceeds what
the residents could remove if all dying and down
material were useable and firewood collection were
restricted to downed wood.
If a selection cutting policy is adopted, minor
disturbances may have to be supplemented with
management tools such as prescribed burning to
prevent natural and dangerous major fires from
destroying the community.
If the management plan contains heavy cuttings
shelterwood, seed tree, or clearcutting (Smith
1962) — mimicking large disturbances on specific
areas, these ares could be spread throughout the
valley and might act as buffers, breaking up the
continuous forests susceptible to fires.
2. The type of firewood cutting will have a major
impact on the volume available and the species
composition of the stands. This will affect both
the firewood supply and the esthetics, two prime
concerns of the National Park Service. Continual
cuttings removing a few trees each time will not
promote young trees in the valley. As shown in
figure 5, no regeneration occurred after small
disturbances. Seedlings appear, but never reach
the 2-in (5.1-cm) diameter class. In fact, field
observations indicate such small openings promote
grasses and shrubby species which may later
exclude new tree stems even if large disturbances
40
30
00
LU
UJ
or
ft 20
cr
10
m
V*
OJ
H HARDWOODS
^ LODGEPOLE PINE
E3 PONDEROSA PINE
□ DOUGLAS -FIR
§
M
i rrm i
It LB
25 55 85 115 145 175 205 235 265
AGE BY 10 -YEAR PERIOD
Figure 6. — Age distributions for all species, all
forest types.
In the past, new stems appeared following major
fires; heavy disturbance, whether natural or man-
made, appeared to be necessary to obtain new stems
again.
3. The type of cutting will have an impact on the
species composition and composition changes within
the valley.
Providing forest fires are avoided, continual
removal of a few stems will change the species com-
position, since the present forests were the
result of relatively large disturbances. It is
probably necessary to mimic such disturbances to
ensure intolerant trees will predominate within
the valley.
If a large forest fire is not avoided and much of
the area is burned, early successional species
(such as cherry, Prunus, will appear dominant for
several decades. It will take 90 years for the
forest to appear again as it does now.
Heavier cutting (perhaps followed by slash burning)
could mimic the natural disturbances and could
be done at regular intervals on scattered areas
throughout the valley. This would result in some
parts of the valley containing vegetation repre-
sentative of different times following a major
disturbance.
4. It is important for forest managers not only
to know the volume and growth rates of the forest
stands but also to know the stand dynamics of each
forest community. Forest inventories unaccom-
panied by knowledge of past disturbances and
present growth rates may lead National Park
Service managers to cut in a manner that would
change the species composition and even the area
of forest. Proper interpretation of forest stand
dynamics may allow managers to manipulate the area
so that management mimics natural growth patterns.
133
LITERATURE CITED
Braun, E. L.
1950. Deciduous forests of eastern North
America. The Blakiston Company, Philadelphia.
596 p.
Daubenmire, R.
19b8. Plant communities: A textbook of plant
synecology. Harper and Row, New York. 300 p.
Dilworth, J. R.
1977. Log scaling and timber cruising. O.S.U.
Book Stores, Inc., Corvallis, Oregon. 473 p.
Meyer, H. A. , and D. D. Stevenson.
1943. The structure and growth of virgin beech-
birch-maple-heralock forests in northern Pennsyl-
vania. J. Agric. Res. 67:465-484.
Minckler, L. S.
1974. Prescribing silvicultural systems.
J. For. 72:269-273.
Oliver, C. D.
1978. The development of northern red oak in
mixed species stands in central New England.
Yale Univ. , Sch. of For. and Envir. Stud. ,
Bull. No. 91.
Donaldson, W. R. , and C. Ruscha.
1975. Washington climate for Chelan, Douglas,
and Okanogan Counties. Coop. Ext. Serv. ,
Washington State Univ. , Pullman, Washington.
Franklin, J. F. , and C. T. Dyrness.
1973. Natural vegetation of Oregon and
Washington. USDA For. Serv. Gen. Tech. Rep.
PNW-8.
Heinselman, M. L.
1973. Fire in the virgin forests of the
Boundary Waters Canoe Area, Minnesota. Quat .
Res. 3:329-382.
Henry, J. D. , and J. M. A. Swan.
1974. Reconstructing forest history from live
and dead plant material — an approach to the
study of forest succession in southwest New
Hampshire. Ecology 55:772-783.
Hough, A. F.
1932. Some diameter distributions in forest
stands of northwestern Pennsylvania. J. For.
30:933-943.
Hough, A. F. , and R. D. Forbes.
1943. The ecology and silvics of forests in the
high plateaus of Pennsylvania. Ecol. Monogr.
13:299-320.
Husch, B., C. I. Miller, and T. W. Beers.
1972. Forest mensuration. Ronald Press.
New York. 410 p.
Oliver, C. D. , and E. P. Stephens.
1977. Reconstruction of a mixed species forest
in central New England. Ecology 58:562-572.
Phillips, E. A.
1959. Methods of vegetation study. Holt,
Rinehart, and Winston, Inc. New York. 107 p.
Smith, David M.
1962. The practice of silviculture,
and Son, Inc., New York. 578 p.
John Wiley
Stubblefield, G. , and C. D. Oliver.
1978. Growth of a mixed red alder-hemlock-
redcedar-Douglas-f ir stand and its silvicul-
tural implications. I_n Utilization and
management of alder. USDA For. Serv. Gen.
Tech. Rep. PNW-70: 307-320.
U.S. Department of Agriculture, Forest Service.
1978. Uneven-aged silviculture and management
in the United States. Timber Manage. Res.
Washington, D. C.
Wierman, C. A. , and C. D. Oliver.
1979. Crown stratification by species in
even-aged Douglas-fir/western hemlock stands of
coastal Washington. Can. J. For. Res. 9:1-9.
Wright, H. E. , Jr.
1974. Landscape development, forest fires, and
wilderness management. Science 186:487-495.
Loucks, 0. L.
1970. Evolution of diversity, efficiency, and
community stability. Amer. Zool. 10:17-25.
134
Optical Properties of Crater Lake, Oregon: Variation in Secchi Disk Transparency,
1937-79
Douglas W. Larson and Mark E. Forbes
ABSTFACT
Since 1937, the Secchi disk transparency of pristine Crater Lake, Oregon, has tended to decrease. Secchi
readings during the summers of 1968 and 1969 averaged 36.6 m which was about 2 m less than the 1937 aver-
age (38.3 m). Recent Secchi measurements (1978 through 1979) averaged 29.3 m (range: 23 to 31 m) which
indicated further decline in lakewater visibility. Cause of the change is uncertain, but an increase in
suspended particulate matter such as phy toplankton is suspected.
Douglas W. Larson, limnologist, Hydrology Section,
U.S. Army Corps of Engineers, Portland District,
Portland, Oregon; and Mark E. Forbes, resource
management specialist, Crater Lake National Park,
USDI, National Park Service, Crater Lake, Oregon.
135
Figure 1. — Secchi disk readings, Crater Lake,
Oregon, 1937-79. Horizontal lines represent the
average values for 20-cm disk readings.
INTRODUCTION
Crater Lake, Oregon, the deepest lake in the United
States at 589 m, is one of the clearest lakes in
the world (Smith and Tyler 1967, Smith et al.
1973). Secchi disk transparency measurements,
indicating the degree of lakewater clarity, were
36, 39, and 40 m in August 1937 (Hasler 1938).
Secchi readings in 1968 and 1969 were slightly less
(average: 36.6 m), although a larger disk measur-
ing 100 cm in diameter was still visible at 44 m
(Larson 1972). Other lakes with exceptionally
high Secchi readings include Lake Masyuko in Japan
(41.6 m, 20-cm disk diameter), Lake Baikal in the
U.S.S.R. (40.6 m, 20-cm disk diameter) and Lake
Tahoe in northern California (36 m, 20-cm disk
diameter) (Thomasson 1956).
Liranological studies were resumed at Crater Lake
in 1978 after a hiatus of nearly 10 years. The
work included Secchi measurements which we present
in this report.
PROCEDURES
Use of the Secchi disk at Crater Lake was under
the most favorable conditions possible (i.e.,
midday readings, clear sky, calm lake surface),
with two and frequently three persons making each
observation. Considerable effort was made to
avoid typical operator errors (Tyler 1968) and to
be consistent when using the disk.
Secchi readings were taken by lowering the disk
(20- and 100-cm diameters) into the lake until it
disappeared. The disk was then raised carefully
until it reappeared. The average distance between
the two sightings was called the Secchi depth
(Welch 1948). Readings were taken only during
summer months.
RESULTS AND DISCUSSION
Secchi measurements of Crater Lake in 1978 and 1979
indicated that lakewater clarity had diminished
since 1969. Whereas 20-cm Secchi readings in 1937
and 1969 averaged 38.3 and 36.6 m respectively, the
average value for 1978-79 was 29.3 m (fig. 1). For
the 100-cra disk, decline in visibility was even
more pronounced — from 44 m on July 16, 1969 (Lar-
son 1972), to 35 and 34 m on July 16 and August 1,
1979, respectively (fig. 1).
136
These values, representing a reduction in lake-
water clarity of about 25 percent since 1937, sug-
gest that the lake has become less transparent due
to an increase, perhaps, in the amount of suspended
particulate matter capable of scattering subsurface
light and reducing Secchi transparency or visibil-
ity. The nature of the particulate matter is not
known (assuming that this is the source of the
problem); but it is possible that the lake has be-
come more productive biologically as the result of
subtle increases in algal nutrients or water tem-
perature. Either factor, singly or in combination,
could conceivably increase phytoplankton biomass,
which is often the cause of reduced Secchi trans-
parency in lakes (Wetzel 1975), and is regarded as
a precursor of accelerated eutrophication in oligo-
trophic systems (Hasler 1969).
The Secchi disk is an inexpensive, easily developed
limnological tool. Nevertheless, its use can re-
veal much about the quality and evolution of a
lake, especially when Secchi measurements span a
period of several years (or decades) and are com-
pared with data from lakes featuring different
levels of eutrophication. The seasonal nature and
sharp numerical fluctuations of lake algae and
other light scattering particulate matter requires,
however, that Secchi measurements be recorded fre-
quently and consistently throughout the year to
avoid inaccurate lake assessments resulting from
sporadic monitoring. Reliable Secchi records, ca-
pable of yielding data for predicting trends, can
usually be obtained even under a financially aus-
tere lake management program. Such a program would
seem to be consistent with National Park Service
goals to preserve those natural resources with
which the agency is charged.
ACKNOWLEDGMENTS
We thank J. Rouse, D. Sholley, B. Wadlington, and
R. Kirschner, Crater Lake National Park, for their
valuable assistance and cooperation. J. Bradley,
Corps of Engineers, and T. Thomas, National Park
Service, assisted in the field.
LITERATURE CITATIONS
Hasler, A. D.
1938. Fish biology and limnology of Crater Lake,
Oregon. J. Wildl. Manage. 2:94-103.
Hasler, A. D.
1969. Cultural eutrophication is reversible.
Bioscience 19:425-431.
Larson, D. W.
1972. Temperature, transparency, and phytoplank-
ton productivity in Crater Lake, Oregon. Limnol.
Oceanogr. 17:410-417.
Smith, R. C. , and J. E. Tyler.
1967. Optical properties of clear natural water.
J. Opt. Soc. Am. 57:589-595.
Smith, R. C, J. E. Tyler, and C. R. Goldman.
1973. Optical properties and color of Lake Tahoe
and Crater Lake. Limnol. Oceanogr. 18:189-199.
Thomasson, K.
1956. Reflections on arctic and alpine lakes.
Oikos 7:117-143.
Tyler, J. E.
1968. The Secchi disc. Limnol. Oceanogr.
13:1-6.
Welch, P. S.
1948. Limnological methods. McGraw-Hill, New
York. 381 p.
Wetzel, R. G.
1975. Limnology. W. B. Saunders, Philadelphia.
743 p.
137
Species Composition and Vertical Distribution of Pelagic Zone Phytoplankton in
Crater Lake, Oregon: 1940-79
Douglas W. Larson and N. Stan Geiger
ABSTRACT
Phytoplankton studies at Crater Lake, Oregon, in 1940 and 1978-79 are compared. Contrary to the earlier
work, (1) diatoms are the predominant phytoplanktonic form, (2) Anabaena sp. is not present, (3) surface
waters contain an abundance of phytoplankton, particularly Nitzschia gracilis, (4) taxa diversity is
greater, with more than 60 species of phytoplankton identified, (5) the vertical distribution of phyto-
plankton varies among species, (6) equally large numbers of organisms occur at various depths throughout
the vertical water column, and (7) phytoplankton smaller than 10 ym appear to be dominant.
Disparity between the two studies might be attributed to either differences in limnological sampling tech-
niques or changes in lake quality.
Douglas W. Larson, 1 imnologi st , Hydrology Section,
U.S. Army Corps of Engineers, Portland District,
Portland, Oregon; and N. Stan Geiger, senior aquatic
biologist, Beak Consultants, Portland, Oregon.
138
INTRODUCTION
MATERIALS AND METHODS
Crater Lake, Oregon, is recognized as one of the
deepest, clearest, most oligotrophic lakes in the
world (Hasler 1938, Byrne 1965, Larson 1972). Yet,
little is known about the biology of this unusual
body of water, including the liranologically import-
ant phy toplankton which occupies the surface-to-
200-m stratum and may extend even deeper.
Brief, cursory surveys of phy toplankton were con-
ducted in 1913 (Kemmerer et al. 1924), 1940 (Utter-
back et al. 1942), and 1959 (Thomasson 1962). More
recent studies (Hoffman and Donaldson 1968, Larson
1972) measured rates of phytoplankton primary pro-
duction and made biomass estimates based on chloro-
phyll a determinations. Here we report on other
aspects of the phytoplankton community, including
taxa identifications and the distribution and
relative abundance of various species through the
vertical water column. The work is intended as a
point of reference for future limnological studies.
PREVIOUS STUDIES
The 1913 plankton survey by Kemmerer et al. (1924)
was probably the first of its kind for the lake.
Although these investigators reported only two
species of phytoplankton, Mougeotia sp. , and a
diatom, Asterionella sp., they discovered that
phytoplankton and zooplankton were distributed to
great depths, reaching maximum abundance between
60 and 200 m.
Limnological studies of the lake in 1940 (Utterback
et al. 1942) indicated that (1) phytoplankton was
most abundant between depths of 75 and 150 m, (2)
virtually no phytoplankton existed in the surface-
to-20-m stratum, or in the deepest sample taken at
425 m, (3) most phytoplankton consisted of filamen-
tous, blue-green algae, Anabaena sp., and (4) dia-
toms constituted only about 15 percent of the total
phytoplankton collected. Samples were obtained by
hauling a No. 20 mesh plankton net (mesh aperture
= 79um) vertically through the water column, or by
casting a Kemmerer bottle to discrete depths,
retrieving the sample, and then centrifuging the
water to extract the phytoplankton. Population
densities reportedly ranged from 1X10-* to 3X10"
cells per liter, but no indication is given as to
the meaning of the term "cells". The work provided
a sketchy taxonomic list, including, in addition
to Anabaena sp., the diatoms Nitzschia sp. , Asteri-
onella sp. , Navicula sp. , and the filamentous alga
Mougeotia sp. identified earlier by Kemmerer et
al. (1924).
Thomasson (1962) listed about a dozen species, one
of which, Ceratium hirundinella, suggested that the
samples were collected from sheltered shoreline lo-
cations rather from the pelagic region of the lake
(sampling locations were not reported). Thomasson
noted that the plankton were "very sparse" on the
day he visited the lake (July 14, 1959), and iden-
tified a rather abundant filamentous alga as
Tribonema sp. which earlier investigators may have
mistakenly called either Mougeotia sp. (Kemmerer
et al. 1924) or Anabaena sp. (Utterback et al.
1942).
The work reported here was completed during the
summers of 1978 and 1979. Water samples were col-
lected with a 2.5-liter (volume) Van Dorn polyvinyl
chloride, messenger-activated bottle from selected
depths extending from lake surface to 200 m. Sam-
ples were retrieved and emptied individually into
an extremely fine mesh ultraplankton net (10- m-
mesh aperture). The collected materials, called
"net" samples, were contained in plastic bottles
and preserved with 3-percent formalin. Filtered
water (about 500 ml) was also collected, preserved
with formalin, and later Millipore-f iltered (HA-
type filters, 0.45-jim pore diameter) to determine
the fraction of phytoplankton small enough to
escape the 10-um net.
The composition of the net samples was first
determined by examining material at various magni-
fications with an inverted microscope (Wild M40).
Samples were then Millipore-f iltered (HA-type) and
examined microscopically (oil immersion, 1000X
magnification). Counts were made of 100 discrete
algal particles (a "particle" was represented by a
single cell, a cluster of cells, or an entire algal
filament) on randomly selected areas of both the
net sample filters and the filters containing the
< 10-um portion of the phytoplankton. Brightfield
and phase microscopy were used for counting and
identification (Keating 1976). Only cells or
groups of cells with distinctive chloroplasts were
considered. Standard taxonomic references for dia-
toms were used (Archibald 1972, Patrick and Reimer
1975) in addition to Sovereign's work (1958) on
Crater Lake periphyton.
Water samples for chlorophyll a_ determinations were
also collected with a Van Dorn bottle and Millipore-
f iltered (HA-type). Extraction and determination
of chlorophyll a_ was done in accordance with
Strickland and Parsons (1965). Percent absorbance
by pigment extracts was measured on a Bausch and
Lomb spectronic spectrophotometer.
RESULTS AND DISCUSSION
The phytoplankton of Crater Lake is dominated by
6 to 10 species, although 63 species have been
identified in 250 of the 350 samples collected to
date (table 1) and 51 species are diatoms, 15 of
which belong to the genus Nitzschia. Some species
Table 1 — Twenty of the more common species of
phytoplankton in Crater Lake, Oregon, 1978-79
Species
Synedra mazamaensis
Nitzschia latens
S. delicatissima
N. frustulum
S. rumpens
N. innominata
S. vaucheriae
N. perminata
Nitzschia gracilis
Stephanodiscus hantzschii
N. demota
S. astraea
N. serpenticula
Asterionella formosa
N. recta
Tribonema CL1
N. mediocris
Gomphonema CL1
N. silica
Melosira italica
139
zo
\ \
40
I /
// JUL* 1*79
• 0
•o
,A
100
Ijs^
l»
^>>
140
^^^
l»0
<^
1*0
\-^
too
^
__ —
1 1 1
'
27 JUNE 1979
to
ft:
LU
Uj
5
I
ft
UJ
Q
I AU6UST 1979
-s
20
■
N. -— -~
^^- - —
29 AUGUST 1979
40
■ — \
•0
■ ' >
V /
•0
I0O
no
wo
^ \
\ >
1*0
180
f r
1 I
200
1 i
'
(4 AUOUST 1979
1 1 10 I » 10
1 1 10 1 1 10*
I • K>=* I
.*?
UNITS £-1
> 10 MM < io AtT\
Figure 1. — Vertical distribution of phytoplankton in Crater Lake, Oregon, 1978-79. Sampling intervals are 20 m.
are periphyton, washed from benthic substrates and
shoreline rocks into the pelagic zone of the lake.
Anabaena sp. was not present. The organism iden-
tified earlier as Anabaena (Utterback et al. 1942)
was probably Tribonema sp. which we found to be
relatively abundant (i.e., averaging 70 percent of
the standing crop at several depths) and similar
in size to the Tribonema described by Thomasson
(i.e., 5-um-wide filaments).
The vertical distribution of phytoplankton in
Crater Lake is characterized, generally, by three
maxima: at the surface (0-20 m) , at middepth (80-
120 m) , and near the bottom of the profile (180-
200 m) (fig. 1). Surface waters usually contain
large numbers of phytoplankton, particularly or-
ganisms smaller than lOum. Ni tzschia gracilis is
by far the most abundant phytoplankter in the
0-20-m stratum (table 2), and its numerical in-
crease through summer appears to be related to an
increase in surface temperatures which reach maxi-
mum levels in late July or early August (Larson
1972). Conversely, phytoplankton in the 180-200-m
stratum is dominated by the small (5-8-jim diameter)
centric diatom Stephanodiscus hantzschii (table 2).
In the middepth region, described earlier as the
lower limit of the photic zone as well as the zone
of maximum productivity (Larson 1972), Tribonema
sp. usually outnumbers all other species present
(table 2). A chlorophyll a maximum also develops
at middepth, but does not correspond well to par-
ticle counts (fig. 2).
140
Table 2 — Relative abundance of the three most common species of phyto-
plankton in Crater Lake, Oregon, 1978-79
Surface
100 m
200 m
Net
fractions
Date
S.h.
N.g.
T.
S.h.
N.g.
T.
S.h.
N.g.
T.
(jm
ercent
1978:
July 11
>10
8
17
37
5
0
83
53
1
35
July 25
> 10
1
77
0
4
0
85
37
4
34
August 7
> 10
0
96
0
9
3
71
18
7
17
< 10
9
89
0
38
2
9
96
1
2
August 29
>10
0
96
0
0
0
88
9
10
22
< 10
2
90
2
10
8
55
87
4
0
1979:
June 27
> 10
1
10
70
1
0
66
7
10
36
< 10
73
19
0
55
25
3
71
12
0
August 1
> 10
0
55
19
2
15
42
5
10
18
August 14
> 10
0
79
8
2
7
54
10
9
22
S.h. = Stephanodiscus
hant
zschii;
N.g-
= Nit
zsch
ii gracilis;
and
T. = Tribonema
CL1.
CHLOROPHYLL a_ (mg m ~3) (Line plot)
O.I 0.2 0.3 0.4 0.5 0.6 0.7 0.8 09 1.0
_ _ ,,,,,,ii,„,f.tiii,,j„,,„,,,i„jii,imiJ„,,,,,rTm
40
60
fiiiiiinlJll, 1,1,11
tiling I \
a.
Q
80
100
120
140
160
l„„l,,,mrTT.
180
200
r,,iii,i,\),j,,ii,iini,iiiii
! 1 1 1 1 r
•""" ,,,,,,.
CHLOROPHYLL £
/ AUGUST 1979
26 AUGUST 1979
ALGAL COUNTS (>I0M>
I AUGUST 1979 I I
14 AUGUST 1979 uiiiiiiiui*
rtmiiiiiizapn v.
11111,1(1,, ,,l A
'in, >,i, in, in
Figure 2. ---Vertical distribution of chlorophyll a
and phytoplankton in Crater Lake, Oregon, 1979.
IxlO4 2xl04 3xl04 4xl04 5 x I04
UNITS J'1 (Histogram)
141
Dissimilar findings between the 1940 and 1978-79
studies are due, possibly, to (1) different samp-
ling and analytical techniques, or (2) alteration
in lake quality, especially with regard to certain
physical and chemical properties capable of influ-
encing the composition, size, and depth distribu-
tion of phy toplankt on populations.
LITERATURE CITATIONS
Archibald, R. E. M.
1972. A preliminary key to the fresh and brack-
ish water species of the genus Nitzschii in
South Africa. News Letter, Limnol. Soc. South
Africa 18:33-46.
The composition and quantity of lake phy toplankton
are often used to characterize the existing state
of lake eut rophication. A continuous, lengthy
record of phy toplankton species types and species
abundance can be valuable in determining the his-
torical condition of the lake as well as the rate
at which the lake is eutrophying. A marked change
in species composition or a growing abundance of a
particular species (or set of species) could be
indicative of a significant shift in lake quality,
perhaps toward an irreversible degraded condition.
Conceivably, through phytoplankton monitoring,
these shifts could be anticipated and averted by
controlling the factors that are potentially harm-
ful to lake environments.
ACKNOWLEDGMENTS
Byrne, J. V.
1965. Morphometry of Crater Lake, Oregon,
nol. Oceanogr. 10:462-465.
Lim-
Hasler, A. D.
1938. Fish biology and limnology of Crater Lake,
Oregon. J. Wildl. Manage. 2:94-103.
Hoffman, F. 0., and J. R. Donaldson.
1968. Zooplankton population dynamics, Crater
Lake, Oregon. Crater Lake Rep. 2. , Natl. Park
Serv. Pub. 12 p.
Keating, K. I.
1976. Algal metabolite influence on bloom se-
quence in eutrophied freshwater ponds. EPA-600/
3-76-081, Environ. Prot. Agency, Washington, D.C.
We thank J. Rouse, M. Forbes, D. Sholley, B. Wad-
lington, and R. Kirschner, Crater Lake National
Park, for their valuable assistance and coopera-
tion. J. Bradley, Corps of Engineers, and T.
Thomas, National Park Service, assisted in the
field. E. Mulvihill, Beak Consultants, kindly
provided laboratory space and equipment. Without
those facilities, this work could not have been
done.
Kemmerer, G. , J. F. Bovard, and W. R. Boorman.
1924. Northwestern lakes of the United States:
Biological and chemical studies with reference
to possibilities in production of fish. U.S.
Bur. Fish. Bull. 39:51-140.
Larson, D. W.
1972. Temperature, transparency, and phytoplank-
ton productivity in Crater Lake, Oregon. Limnol.
Oceanogr. 17:410-417.
Patrick, R. , and C. Reimer.
1975. The diatoms of the United States. Exclu-
sive of Alaska and Hawaii. Vol. 2, Part 1.
Acad. Nat. Soc, No. 13, Philadelphia.
Sovereign, H. E.
1958. The diatoms of Crater Lake, Oregon.
Trans. Am. Micros. Soc. 77:96-124.
Strickland, J. D. H. , and T. R. Parsons.
1965. A manual of seawater analysis. Fish. Res.
Bd. Can. Bull. 125.
Thomasson, K.
1962. Planktological notes from western North
America. Arkiv fur Botanik 4:437-463.
Utterback, C. L. , L. D. Phifer, and R. J. Robinson.
1942. Some planktonic and optical characteris-
tics of Crater Lake. Ecology 23:97-103.
142
QH 104.5.N6C66 1979, 00011
3 1604 002 272 328
QEL0U.5
.N6C66
1979
61+01*65
myr®
87
n
Live Music
Archive
Librivox
Free Audio
Metropolitan Museum
Cleveland
Museum of Art
Internet
Arcade
Console Living Room
Open Library
American
Libraries
TV News
Understanding
9/11